Publications | IDEAS Lab

2025

CVPR

Split Adaptation for Pre-trained Vision Transformers

Lixu Wang, Bingqi Shang, Yi Li, and 4 more authors

In 2025 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2025

@inproceedings{wang_split_2025,
  booktitle = {2025 {IEEE} {Conference} on {Computer} {Vision} and {Pattern} {Recognition} ({CVPR})},
  title = {Split {Adaptation} for {Pre}-trained {Vision} {Transformers}},
  url = {https://openaccess.thecvf.com/content/CVPR2025/html/Wang_Split_Adaptation_for_Pre-trained_Vision_Transformers_CVPR_2025_paper.html},
  language = {en},
  urldate = {2025-06-18},
  author = {Wang, Lixu and Shang, Bingqi and Li, Yi and Mohapatra, Payal and Dong, Wei and Wang, Xiao and Zhu, Qi},
  year = {2025},
  pages = {20092--20102}
}

ACL
Can LLMs Understand Unvoiced Speech? Exploring EMG-to-Text Conversion with LLMs

Payal Mohapatra, Akash Pandey, Xiaoyuan Zhang, and 1 more author

In Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 2: Short Papers), Jul 2025

Abs DOI Bib

Unvoiced electromyography (EMG) is an effective communication tool for individuals unable to produce vocal speech. However, most prior methods rely on paired voiced and unvoiced EMG signals, along with speech data, for unvoiced EMG-to-text conversion, which is not practical for these individuals. Given the rise of large language models (LLMs) in speech recognition, we explore their potential to understand unvoiced speech. To this end, we address the challenge of \textitlearning from unvoiced EMG alone and propose a novel EMG adaptor module that maps EMG features to an LLM’s input space, achieving an average word error rate of 0.49 on a closed-vocabulary unvoiced EMG-to-text task. Even with a conservative data availability of just six minutes, our approach improves performance over specialized models by nearly 20%. While LLMs have been shown to be extendable to new language modalities—such as audio—understanding articulatory biosignals, like unvoiced EMG, is more challenging. This work takes a crucial first step toward enabling LLMs to comprehend unvoiced speech using surface EMG.
@inproceedings{mohapatra-etal-2025-llms, title = {Can {LLM}s Understand Unvoiced Speech? Exploring {EMG}-to-Text Conversion with {LLM}s}, author = {Mohapatra, Payal and Pandey, Akash and Zhang, Xiaoyuan and Zhu, Qi}, editor = {Che, Wanxiang and Nabende, Joyce and Shutova, Ekaterina and Pilehvar, Mohammad Taher}, booktitle = {Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 2: Short Papers)}, month = jul, year = {2025}, address = {Vienna, Austria}, publisher = {Association for Computational Linguistics}, url = {https://aclanthology.org/2025.acl-short.56/}, doi = {10.18653/v1/2025.acl-short.56}, pages = {703--712}, isbn = {979-8-89176-252-7}, keywords = {Computer Science - Computation and Language} }

2024

Interspeech

Missingness-resilient Video-enhanced Multimodal Disfluency Detection

Payal Mohapatra, Shamika Likhite, Subrata Biswas, and 2 more authors

In Interspeech 2024, Sep 2024

DOI Bib HTML

@inproceedings{mohapatra_missingness-resilient_2024,
  title = {Missingness-resilient {Video}-enhanced {Multimodal} {Disfluency} {Detection}},
  url = {https://www.isca-archive.org/interspeech_2024/mohapatra24_interspeech.html},
  doi = {10.21437/Interspeech.2024-1458},
  language = {en},
  urldate = {2025-06-18},
  booktitle = {Interspeech 2024},
  publisher = {ISCA},
  author = {Mohapatra, Payal and Likhite, Shamika and Biswas, Subrata and Islam, Bashima and Zhu, Qi},
  month = sep,
  year = {2024},
  pages = {5093--5097}
}

PNAS Nexus
Wearable network for multilevel physical fatigue prediction in manufacturing workers

Payal Mohapatra, Vasudev Aravind, Marisa Bisram, and 13 more authors

PNAS Nexus, Oct 2024

Abs DOI Bib HTML

Manufacturing workers face prolonged strenuous physical activities, impacting both financial aspects and their health due to work-related fatigue. Continuously monitoring physical fatigue and providing meaningful feedback is crucial to mitigating human and monetary losses in manufacturing workplaces. This study introduces a novel application of multimodal wearable sensors and machine learning techniques to quantify physical fatigue and tackle the challenges of real-time monitoring on the factory floor. Unlike past studies that view fatigue as a dichotomous variable, our central formulation revolves around the ability to predict multilevel fatigue, providing a more nuanced understanding of the subject’s physical state. Our multimodal sensing framework is designed for continuous monitoring of vital signs, including heart rate, heart rate variability, skin temperature, and more, as well as locomotive signs by employing inertial motion units strategically placed at six locations on the upper body. This comprehensive sensor placement allows us to capture detailed data from both the torso and arms, surpassing the capabilities of single-point data collection methods. We developed an innovative asymmetric loss function for our machine learning model, which enhances prediction accuracy for numerical fatigue levels and supports real-time inference. We collected data on 43 subjects following an authentic manufacturing protocol and logged their self-reported fatigue. Based on the analysis, we provide insights into our multilevel fatigue monitoring system and discuss results from an in-the-wild evaluation of actual operators on the factory floor. This study demonstrates our system’s practical applicability and contributes a valuable open-access database for future research.
@article{mohapatra_wearable_2024, title = {Wearable network for multilevel physical fatigue prediction in manufacturing workers}, volume = {3}, copyright = {https://creativecommons.org/licenses/by-nc/4.0/}, issn = {2752-6542}, url = {https://academic.oup.com/pnasnexus/article/doi/10.1093/pnasnexus/pgae421/7815440}, doi = {10.1093/pnasnexus/pgae421}, language = {en}, number = {10}, urldate = {2025-06-18}, journal = {PNAS Nexus}, author = {Mohapatra, Payal and Aravind, Vasudev and Bisram, Marisa and Lee, Young-Joong and Jeong, Hyoyoung and Jinkins, Katherine and Gardner, Richard and Streamer, Jill and Bowers, Brent and Cavuoto, Lora and Banks, Anthony and Xu, Shuai and Rogers, John and Cao, Jian and Zhu, Qi and Guo, Ping}, editor = {Yue, Xiaowei}, month = oct, year = {2024}, pages = {pgae421} }
J. Manuf. Syst.
Towards a digital twin framework in additive manufacturing: Machine learning and bayesian optimization for time series process optimization

Vispi Karkaria, Anthony Goeckner, Rujing Zha, and 6 more authors

Journal of Manufacturing Systems, May 2024

Abs DOI Bib HTML

Laser directed-energy deposition (DED) offers notable advantages in additive manufacturing (AM) for producing intricate geometries and facilitating material functional grading. However, inherent challenges such as material property inconsistencies and part variability persist, predominantly due to its layer-wise fabrication approach. Critical to these challenges is heat accumulation during DED, influencing the resultant material microstructure and properties. Although closed-loop control methods for managing heat accumulation and temperature regulation are prevalent in DED literature, few approaches integrate real-time monitoring, physics-based modeling, and control simultaneously in a cohesive framework. To address this, we present a digital twin (DT) framework for real-time model predictive control of process parameters of the DED for achieving a specific process design objective. To enable its implementation, we detail the development of a surrogate model utilizing Long Short-Term Memory (LSTM)-based machine learning which uses Bayesian Inference to predict temperatures across various spatial locations of the DED-built part. This model offers real-time predictions of future temperature states. In addition, we introduce a Bayesian Optimization (BO) method for Time Series Process Optimization (BOTSPO). Its foundational principles align with traditional BO, and its novelty lies in our unique time series process profile generator with a reduced dimensional representation. BOTSPO is used for dynamic process optimization in which we deploy BOTSPO to determine the optimal laser power profile, aiming to achieve desired mechanical properties in a DED build. The identified profile establishes a process trajectory that online process optimizations aim to match or exceed in performance. This paper elucidates components of the digital twin framework, advocating its prospective consolidation into a comprehensive digital twin system for AM.
@article{karkaria_towards_2024, title = {Towards a digital twin framework in additive manufacturing: {Machine} learning and bayesian optimization for time series process optimization}, copyright = {All rights reserved}, issn = {0278-6125}, shorttitle = {Towards a digital twin framework in additive manufacturing}, url = {https://www.sciencedirect.com/science/article/pii/S027861252400089X}, doi = {10.1016/j.jmsy.2024.04.023}, urldate = {2024-05-08}, journal = {Journal of Manufacturing Systems}, author = {Karkaria, Vispi and Goeckner, Anthony and Zha, Rujing and Chen, Jie and Zhang, Jianjing and Zhu, Qi and Cao, Jian and Gao, Robert X. and Chen, Wei}, month = may, year = {2024}, keywords = {Digital twin, Additive manufacturing, Recurrent neural network, Directed energy deposition, Process optimization, Bayesian optimization, Long-short term memory} }

DAC

Invited: Algorithm and Hardware Co-Design for Energy-Efficient Neural SLAM

Lingyi Huang, Cheng Yang, Yu Gong, and 5 more authors

In Proceedings of the 61st ACM/IEEE Design Automation Conference, Nov 2024

Abs DOI Bib HTML

In this paper, we introduce a novel approach to enhancing neural network-based Simultaneous Localization and Mapping (SLAM) through the integration of model compression techniques and customized hardware architecture that focuses on micro-architectural and dataflow optimizations to improve computational efficiency and performance. Experiments across different scenarios demonstrate that the proposed approach achieves significant improvement.

@inproceedings{huang_invited_2024,
  address = {New York, NY, USA},
  series = {{DAC} '24},
  title = {Invited: {Algorithm} and {Hardware} {Co}-{Design} for {Energy}-{Efficient} {Neural} {SLAM}},
  isbn = {979-8-4007-0601-1},
  shorttitle = {Invited},
  url = {https://dl.acm.org/doi/10.1145/3649329.3664829},
  doi = {10.1145/3649329.3664829},
  urldate = {2024-11-08},
  booktitle = {Proceedings of the 61st {ACM}/{IEEE} {Design} {Automation} {Conference}},
  publisher = {Association for Computing Machinery},
  author = {Huang, Lingyi and Yang, Cheng and Gong, Yu and Sui, Yang and Zang, Xiao and Goeckner, Anthony and Zhu, Qi and Yuan, Bo},
  month = nov,
  year = {2024},
  pages = {1--4}
}

IROS
Graph Neural Network-based Multi-agent Reinforcement Learning for Resilient Distributed Coordination of Multi-Robot Systems

Anthony Goeckner, Yueyuan Sui, Nicolas Martinet, and 2 more authors

In 2024 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), Oct 2024

ISSN: 2153-0866 code: https://github.com/NU-IDEAS-Lab/patrolling_zoo

Abs DOI Bib HTML

Existing multi-agent coordination techniques are often fragile and vulnerable to anomalies such as agent attrition and communication disturbances, which are quite common in the real-world deployment of systems like field robotics. To better prepare these systems for the real world, we present a graph neural network (GNN)-based multi-agent reinforcement learning (MARL) method for resilient distributed coordination of a multi-robot system. Our method, Multi-Agent Graph Embedding-based Coordination (MAGEC), is trained using multi-agent proximal policy optimization (PPO) and enables distributed coordination around global objectives under agent attrition, partial observability, and limited or disturbed communications. We use a multi-robot patrolling scenario to demonstrate our MAGEC method in a ROS 2-based simulator and then compare its performance with prior coordination approaches. Results demonstrate that MAGEC outperforms existing methods in several experiments involving agent attrition and communication disturbance, and provides competitive results in scenarios without such anomalies.
@inproceedings{goeckner_graph_2024, title = {Graph {Neural} {Network}-based {Multi}-agent {Reinforcement} {Learning} for {Resilient} {Distributed} {Coordination} of {Multi}-{Robot} {Systems}}, url = {https://ieeexplore.ieee.org/document/10802510}, doi = {10.1109/IROS58592.2024.10802510}, urldate = {2025-01-03}, booktitle = {2024 {IEEE}/{RSJ} {International} {Conference} on {Intelligent} {Robots} and {Systems} ({IROS})}, author = {Goeckner, Anthony and Sui, Yueyuan and Martinet, Nicolas and Li, Xinliang and Zhu, Qi}, month = oct, year = {2024}, note = {ISSN: 2153-0866 code: https://github.com/NU-IDEAS-Lab/patrolling\_zoo}, keywords = {Graph neural networks, Intelligent robots, Multi-robot systems, Navigation, Observability, Optimization, Reinforcement learning, Robot kinematics, Solids, Vehicle routing}, pages = {5732--5739} }
RA-L
Attrition-Aware Adaptation for Multi-Agent Patrolling

Anthony Goeckner, Xinliang Li, Ermin Wei, and 1 more author

IEEE Robotics and Automation Letters, Aug 2024

Conference Name: IEEE Robotics and Automation Letters code: https://github.com/NU-IDEAS-Lab/patrolling_sim

Abs DOI Bib HTML

Multi-agent patrolling is a key problem in a variety of domains such as intrusion detection, area surveillance, and policing, which involves repeated visits by a group of agents to specified points in an environment. While the problem is well-studied, most works do not provide performance guarantees and either do not consider agent attrition or impose significant communication requirements to enable adaptation. In this work, we present the Adaptive Heuristic-based Patrolling Algorithm, which is capable of adaptation to agent loss using minimal communication by taking advantage of Voronoi partitioning, and which meets guaranteed performance bounds. Additionally, we provide new centralized and distributed mathematical programming formulations of the patrolling problem, analyze the properties of Voronoi partitioning, and finally, show the value of our adaptive heuristic algorithm by comparison with various benchmark algorithms using physical robots and simulation based on the Robot Operating System (ROS) 2.
@article{goeckner_attrition-aware_2024, title = {Attrition-{Aware} {Adaptation} for {Multi}-{Agent} {Patrolling}}, volume = {9}, copyright = {All rights reserved}, issn = {2377-3766}, url = {https://ieeexplore.ieee.org/document/10582411?source=authoralert}, doi = {10.1109/LRA.2024.3421793}, number = {8}, urldate = {2024-07-10}, journal = {IEEE Robotics and Automation Letters}, author = {Goeckner, Anthony and Li, Xinliang and Wei, Ermin and Zhu, Qi}, month = aug, year = {2024}, note = {Conference Name: IEEE Robotics and Automation Letters code: https://github.com/NU-IDEAS-Lab/patrolling\_sim}, keywords = {Electric breakdown, Heuristic algorithms, Multi-robot systems, Partitioning algorithms, path planning for multiple mobile robots or agents, Resource management, robotics in hazardous fields, Task analysis, Vectors, Vehicle routing}, pages = {7230--7237} }

L4DC

State-wise safe reinforcement learning with pixel observations

Sinong Zhan, Yixuan Wang, Qingyuan Wu, and 3 more authors

In 6th Annual Learning for Dynamics & Control Conference, 2024

Bib

@inproceedings{zhan2024state,
  title = {State-wise safe reinforcement learning with pixel observations},
  author = {Zhan, Sinong and Wang, Yixuan and Wu, Qingyuan and Jiao, Ruochen and Huang, Chao and Zhu, Qi},
  booktitle = {6th Annual Learning for Dynamics \& Control Conference},
  pages = {1187--1201},
  year = {2024},
  organization = {PMLR},
}

ICML

Boosting Reinforcement Learning with Strongly Delayed Feedback Through Auxiliary Short Delays

Qingyuan Wu, Simon Sinong Zhan, Yixuan Wang, and 6 more authors

In International Conference on Machine Learning, 2024

Bib

@inproceedings{wu2024boosting,
  title = {Boosting Reinforcement Learning with Strongly Delayed Feedback Through Auxiliary Short Delays},
  author = {Wu, Qingyuan and Zhan, Simon Sinong and Wang, Yixuan and Wang, Yuhui and Lin, Chung-Wei and Lv, Chen and Zhu, Qi and Schmidhuber, J{\"u}rgen and Huang, Chao},
  booktitle = {International Conference on Machine Learning},
  pages = {53973--53998},
  year = {2024},
  organization = {PMLR}
}

NeurIPS

Variational delayed policy optimization

Qingyuan Wu^*, Simon S Zhan^*, Yixuan Wang, and 5 more authors

Advances in neural information processing systems, 2024

Bib

@article{wu2024variational,
  title = {Variational delayed policy optimization},
  author = {Wu, Qingyuan and Zhan, Simon S and Wang, Yixuan and Wang, Yuhui and Lin, Chung-Wei and Lv, Chen and Zhu, Qi and Huang, Chao},
  journal = {Advances in neural information processing systems},
  volume = {37},
  pages = {54330--54356},
  year = {2024}
}

IROS

Kinematics-aware trajectory generation and prediction with latent stochastic differential modeling

Ruochen Jiao, Yixuan Wang, Xiangguo Liu, and 3 more authors

In 2024 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), 2024

Bib

@inproceedings{jiao2024kinematics,
  title = {Kinematics-aware trajectory generation and prediction with latent stochastic differential modeling},
  author = {Jiao, Ruochen and Wang, Yixuan and Liu, Xiangguo and Zhan, Simon Sinong and Huang, Chao and Zhu, Qi},
  booktitle = {2024 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)},
  pages = {565--572},
  year = {2024},
  organization = {IEEE}
}

Case study: runtime safety verification of neural network controlled system

Frank Yang, Sinong Simon Zhan, Yixuan Wang, and 2 more authors

In International Conference on Runtime Verification, 2024

Bib

@inproceedings{yang2024case,
  title = {Case study: runtime safety verification of neural network controlled system},
  author = {Yang, Frank and Zhan, Sinong Simon and Wang, Yixuan and Huang, Chao and Zhu, Qi},
  booktitle = {International Conference on Runtime Verification},
  pages = {205--217},
  year = {2024},
  organization = {Springer}
}

ICASSP

DACR: Distribution-Augmented Contrastive Reconstruction for Time-Series Anomaly Detection

Lixu Wang, Shichao Xu, Xinyu Du, and 1 more author

In ICASSP 2024-2024 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), 2024

Bib

@inproceedings{wang2024dacr,
  title = {DACR: Distribution-Augmented Contrastive Reconstruction for Time-Series Anomaly Detection},
  author = {Wang, Lixu and Xu, Shichao and Du, Xinyu and Zhu, Qi},
  booktitle = {ICASSP 2024-2024 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)},
  pages = {7545--7549},
  year = {2024},
  organization = {IEEE}
}

NeurIPS

Semantic feature learning for universal unsupervised cross-domain retrieval

Lixu Wang, Xinyu Du, and Qi Zhu

Advances in Neural Information Processing Systems, 2024

Bib

@article{wang2024semantic,
  title = {Semantic feature learning for universal unsupervised cross-domain retrieval},
  author = {Wang, Lixu and Du, Xinyu and Zhu, Qi},
  journal = {Advances in Neural Information Processing Systems},
  volume = {37},
  pages = {79516--79539},
  year = {2024}
}

2023

Joint Differentiable Optimization and Verification for Certified Reinforcement Learning

Yixuan Wang, Simon Zhan, Zhilu Wang, and 4 more authors

Mar 2023

arXiv:2201.12243 [cs, eess]

Abs DOI Bib HTML

In model-based reinforcement learning for safety-critical control systems, it is important to formally certify system properties (e.g., safety, stability) under the learned controller. However, as existing methods typically apply formal verification }emph{after} the controller has been learned, it is sometimes difficult to obtain any certificate, even after many iterations between learning and verification. To address this challenge, we propose a framework that jointly conducts reinforcement learning and formal verification by formulating and solving a novel bilevel optimization problem, which is differentiable by the gradients from the value function and certificates. Experiments on a variety of examples demonstrate the significant advantages of our framework over the model-based stochastic value gradient (SVG) method and the model-free proximal policy optimization (PPO) method in finding feasible controllers with barrier functions and Lyapunov functions that ensure system safety and stability.
@misc{wangJointDifferentiableOptimization2023, title = {Joint {Differentiable} {Optimization} and {Verification} for {Certified} {Reinforcement} {Learning}}, url = {http://arxiv.org/abs/2201.12243}, doi = {10.48550/arXiv.2201.12243}, urldate = {2023-04-04}, publisher = {arXiv}, author = {Wang, Yixuan and Zhan, Simon and Wang, Zhilu and Huang, Chao and Wang, Zhaoran and Yang, Zhuoran and Zhu, Qi}, month = mar, year = {2023}, note = {arXiv:2201.12243 [cs, eess]}, keywords = {Computer Science - Machine Learning, Electrical Engineering and Systems Science - Systems and Control} }
Connectivity Enhanced Safe Neural Network Planner for Lane Changing in Mixed Traffic

Xiangguo Liu, Ruochen Jiao, Bowen Zheng, and 2 more authors

Feb 2023

arXiv:2302.02513 [cs]

Abs DOI Bib HTML

Connectivity technology has shown great potentials in improving the safety and efficiency of transportation systems by providing information beyond the perception and prediction capabilities of individual vehicles. However, it is expected that human-driven and autonomous vehicles, and connected and non-connected vehicles need to share the transportation network during the transition period to fully connected and automated transportation systems. Such mixed traffic scenarios significantly increase the complexity in analyzing system behavior and quantifying uncertainty for highly interactive scenarios, e.g., lane changing. It is even harder to ensure system safety when neural network based planners are leveraged to further improve efficiency. In this work, we propose a connectivity-enhanced neural network based lane changing planner. By cooperating with surrounding connected vehicles in dynamic environment, our proposed planner will adapt its planned trajectory according to the analysis of a safe evasion trajectory. We demonstrate the strength of our planner design in improving efficiency and ensuring safety in various mixed traffic scenarios with extensive simulations. We also analyze the system robustness when the communication or coordination is not perfect.
@misc{liuConnectivityEnhancedSafe2023, title = {Connectivity {Enhanced} {Safe} {Neural} {Network} {Planner} for {Lane} {Changing} in {Mixed} {Traffic}}, url = {http://arxiv.org/abs/2302.02513}, doi = {10.48550/arXiv.2302.02513}, urldate = {2023-04-04}, publisher = {arXiv}, author = {Liu, Xiangguo and Jiao, Ruochen and Zheng, Bowen and Liang, Dave and Zhu, Qi}, month = feb, year = {2023}, note = {arXiv:2302.02513 [cs]}, keywords = {Computer Science - Robotics} }
Mixed-Traffic Intersection Management Utilizing Connected and Autonomous Vehicles as Traffic Regulators

Pin-Chun Chen, Xiangguo Liu, Chung-Wei Lin, and 2 more authors

In Proceedings of the 28th Asia and South Pacific Design Automation Conference, Jan 2023

Abs DOI Bib HTML

Connected and autonomous vehicles (CAVs) can realize many revolutionary applications, but it is expected to have mixed-traffic including CAVs and human-driving vehicles (HVs) together for decades. In this paper, we target the problem of mixed-traffic intersection management and schedule CAVs to control the subsequent HVs. We develop a dynamic programming approach and a mixed integer linear programming (MILP) formulation to optimally solve the problems with the corresponding intersection models. We then propose an MILPbased approach which is more efficient and real-time-applicable than solving the optimal MILP formulation, while keeping good solution quality as well as outperforming the first-come-first-served (FCFS) approach. Experimental results and SUMO simulation indicate that controlling CAVs by our approaches is effective to regulate mixedtraffic even if the CAV penetration rate is low, which brings incentive to early adoption of CAVs.
@inproceedings{chenMixedTrafficIntersectionManagement2023, address = {Tokyo Japan}, title = {Mixed-{Traffic} {Intersection} {Management} {Utilizing} {Connected} and {Autonomous} {Vehicles} as {Traffic} {Regulators}}, isbn = {978-1-4503-9783-4}, url = {https://dl.acm.org/doi/10.1145/3566097.3567849}, doi = {10.1145/3566097.3567849}, language = {en}, urldate = {2023-04-04}, booktitle = {Proceedings of the 28th {Asia} and {South} {Pacific} {Design} {Automation} {Conference}}, publisher = {ACM}, author = {Chen, Pin-Chun and Liu, Xiangguo and Lin, Chung-Wei and Huang, Chao and Zhu, Qi}, month = jan, year = {2023}, pages = {52--57} }
Safety-Driven Interactive Planning for Neural Network-Based Lane Changing

Xiangguo Liu, Ruochen Jiao, Bowen Zheng, and 2 more authors

In Proceedings of the 28th Asia and South Pacific Design Automation Conference, Jan 2023

Abs DOI Bib HTML

Neural network-based driving planners have shown great promises in improving task performance of autonomous driving. However, it is critical and yet very challenging to ensure the safety of systems with neural network-based components, especially in dense and highly interactive traffic environments. In this work, we propose a safety-driven interactive planning framework for neural networkbased lane changing. To prevent over-conservative planning, we identify the driving behavior of surrounding vehicles and assess their aggressiveness, and then adapt the planned trajectory for the ego vehicle accordingly in an interactive manner. The ego vehicle can proceed to change lanes if a safe evasion trajectory exists even in the predicted worst case; otherwise, it can stay around the current lateral position or return back to the original lane. We quantitatively demonstrate the effectiveness of our planner design and its advantage over baseline methods through extensive simulations with diverse and comprehensive experimental settings, as well as in real-world scenarios collected by an autonomous vehicle company.
@inproceedings{liuSafetyDrivenInteractivePlanning2023, address = {Tokyo Japan}, title = {Safety-{Driven} {Interactive} {Planning} for {Neural} {Network}-{Based} {Lane} {Changing}}, isbn = {978-1-4503-9783-4}, url = {https://dl.acm.org/doi/10.1145/3566097.3567847}, doi = {10.1145/3566097.3567847}, language = {en}, urldate = {2023-04-04}, booktitle = {Proceedings of the 28th {Asia} and {South} {Pacific} {Design} {Automation} {Conference}}, publisher = {ACM}, author = {Liu, Xiangguo and Jiao, Ruochen and Zheng, Bowen and Liang, Dave and Zhu, Qi}, month = jan, year = {2023}, pages = {39--45} }
Deja Vu: Continual Model Generalization for Unseen Domains

Chenxi Liu, Lixu Wang, Lingjuan Lyu, and 3 more authors

In , Feb 2023

Abs Bib HTML

In real-world applications, deep learning models often run in non-stationary environments where the target data distribution continually shifts over time. There have been numerous domain adaptation (DA) methods in both online and offline modes to improve cross-domain adaptation ability. However, these DA methods typically only provide good performance after a long period of adaptation, and perform poorly on new domains before and during adaptation – in what we call the “Unfamiliar Period”, especially when domain shifts happen suddenly and significantly. On the other hand, domain generalization (DG) methods have been proposed to improve the model generalization ability on unadapted domains. However, existing DG works are ineffective for continually changing domains due to severe catastrophic forgetting of learned knowledge. To overcome these limitations of DA and DG in handling the Unfamiliar Period during continual domain shift, we propose RaTP, a framework that focuses on improving models’ target domain generalization (TDG) capability, while also achieving effective target domain adaptation (TDA) capability right after training on certain domains and forgetting alleviation (FA) capability on past domains. RaTP includes a training-free data augmentation module to prepare data for TDG, a novel pseudo-labeling mechanism to provide reliable supervision for TDA, and a prototype contrastive alignment algorithm to align different domains for achieving TDG, TDA and FA. Extensive experiments on Digits, PACS, and DomainNet demonstrate that RaTP significantly outperforms state-of-the-art works from Continual DA, Source-Free DA, Test-Time/Online DA, Single DG, Multiple DG and Unified DA&DG in TDG, and achieves comparable TDA and FA capabilities.
@inproceedings{liuDejaVuContinual2023, title = {Deja {Vu}: {Continual} {Model} {Generalization} for {Unseen} {Domains}}, shorttitle = {Deja {Vu}}, url = {https://openreview.net/forum?id=L8iZdgeKmI6}, language = {en}, urldate = {2023-04-04}, author = {Liu, Chenxi and Wang, Lixu and Lyu, Lingjuan and Sun, Chen and Wang, Xiao and Zhu, Qi}, month = feb, year = {2023} }
A Safety-Guaranteed Framework for Neural-Network-Based Planners in Connected Vehicles under Communication Disturbance

Kevin Kai-Chun Chang, Xiangguo Liu, Chung-Wei Lin, and 2 more authors

In 2023 Design Automation and Test in Europe Conference, Apr 2023

Abs Bib

Neural-network-based (NN-based) planners have been increasingly used to enhance the performance of planning for autonomous vehicles. However, it is often difficult for NN-based planners to balance efficiency and safety in complicated scenarios, especially under real-world communication disturbance. To tackle this challenge, we present a safety-guaranteed framework for NNbased planners in connected vehicle environments with communication disturbance. Given any NN-based planner with no safetyguarantee, the framework generates a robust compound planner embedding the NN-based planner to ensure overall system safety. Moreover, with the aid of an information filter for imperfect communication and an aggressive approach for the estimation of the unsafe set, the compound planner could achieve similar or better efficiency than the given NN-based planner. A comprehensive case study of unprotected left turn and extensive simulations demonstrate the effectiveness of our framework.
@inproceedings{changSafetyGuaranteedFrameworkNeuralNetworkBased2023, title = {A {Safety}-{Guaranteed} {Framework} for {Neural}-{Network}-{Based} {Planners} in {Connected} {Vehicles} under {Communication} {Disturbance}}, language = {en}, booktitle = {2023 {Design} {Automation} and {Test} in {Europe} {Conference}}, author = {Chang, Kevin Kai-Chun and Liu, Xiangguo and Lin, Chung-Wei and Huang, Chao and Zhu, Qi}, month = apr, year = {2023} }
ICASSP
Efficient Stuttering Event Detection Using Siamese Networks

Payal Mohapatra, Bashima Islam, Md Tamzeed Islam, and 2 more authors

In ICASSP 2023 - 2023 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), Jun 2023

ISSN: 2379-190X

Abs DOI Bib HTML

Speech disfluency research is pivotal to accommodating atypical speakers in mainstream conversational technology. However, the lack of publicly available labeled and unlabeled datasets is a significant bottleneck to such research. While many works use pseudo dysfluency data with proxy labels and formulate a self-supervised task, we see merit in using real-world data. In this work, we consolidate the corpora of publicly available speech disfluency datasets with and without labels and propose DisfluentSiam – an efficient siamese network-based small-scale pretraining pipeline using task-specific data from multiple domains with only 10M trainable parameters. We show that with DisfluentSiam, we achieve an average of 15% boost in performance across five types of dysfluency event detection compared to direct wav2vec 2.0 embeddings. In particular, with only 4-5 mins of labeled data for fine-tuning, the DisfluentSiam demonstrates the advantage of task-specific pretraining with up to 25% higher accuracy.
@inproceedings{mohapatra_efficient_2023, title = {Efficient {Stuttering} {Event} {Detection} {Using} {Siamese} {Networks}}, url = {https://ieeexplore.ieee.org/abstract/document/10094692}, doi = {10.1109/ICASSP49357.2023.10094692}, urldate = {2025-06-18}, booktitle = {{ICASSP} 2023 - 2023 {IEEE} {International} {Conference} on {Acoustics}, {Speech} and {Signal} {Processing} ({ICASSP})}, author = {Mohapatra, Payal and Islam, Bashima and Islam, Md Tamzeed and Jiao, Ruochen and Zhu, Qi}, month = jun, year = {2023}, note = {ISSN: 2379-190X}, keywords = {Data models, Real-time systems, Signal processing, Acoustics, Dysfluency, Event detection, Machine translation, Pipelines, Self-supervised Learning}, pages = {1--5} }
ICASSP
Person Identification with Wearable Sensing Using Missing Feature Encoding and Multi-Stage Modality Fusion

Payal Mohapatra, Akash Pandey, Sinan Keten, and 2 more authors

In ICASSP 2023 - 2023 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), Jun 2023

ISSN: 2379-190X

Abs DOI Bib HTML

We present a missingness-aware fusion network (MAFN) to identify a person’s digital phenotype from continuously measured longitudinal multi-modal wearable data. This work is done as a part of Track 1 of e-Prevention: Person Identification and Relapse Detection from Continuous Recordings of Biosignals Signal Processing Grand Challenge at International Conference on Acoustics, Speech, & Signal Processing (ICASSP) 2023. MAFN achieves an accuracy of 91.36% on test data. Additionally, our experiments confirm findings from previous works that kinetic features derived from the accelerometer in-deed contain more discriminative features for person identification task.
@inproceedings{mohapatra_person_2023, title = {Person {Identification} with {Wearable} {Sensing} {Using} {Missing} {Feature} {Encoding} and {Multi}-{Stage} {Modality} {Fusion}}, url = {https://ieeexplore.ieee.org/document/10097005/}, doi = {10.1109/ICASSP49357.2023.10097005}, urldate = {2025-06-18}, booktitle = {{ICASSP} 2023 - 2023 {IEEE} {International} {Conference} on {Acoustics}, {Speech} and {Signal} {Processing} ({ICASSP})}, author = {Mohapatra, Payal and Pandey, Akash and Keten, Sinan and Chen, Wei and Zhu, Qi}, month = jun, year = {2023}, note = {ISSN: 2379-190X}, keywords = {Sensors, Encoding, Acoustics, Accelerometers, Acoustic measurements, Kinetic theory, Recording}, pages = {1--2} }

Effect of Attention and Self-Supervised Speech Embeddings on Non-Semantic Speech Tasks

Payal Mohapatra, Akash Pandey, Yueyuan Sui, and 1 more author

In Proceedings of the 31st ACM International Conference on Multimedia, Oct 2023

DOI Bib HTML

@inproceedings{mohapatra_effect_2023,
  address = {Ottawa ON Canada},
  title = {Effect of {Attention} and {Self}-{Supervised} {Speech} {Embeddings} on {Non}-{Semantic} {Speech} {Tasks}},
  isbn = {979-8-4007-0108-5},
  url = {https://dl.acm.org/doi/10.1145/3581783.3612855},
  doi = {10.1145/3581783.3612855},
  language = {en},
  urldate = {2025-06-18},
  booktitle = {Proceedings of the 31st {ACM} {International} {Conference} on {Multimedia}},
  publisher = {ACM},
  author = {Mohapatra, Payal and Pandey, Akash and Sui, Yueyuan and Zhu, Qi},
  month = oct,
  year = {2023},
  pages = {9511--9515}
}

DAC
Invited: Waving the Double-Edged Sword: Building Resilient CAVs with Edge and Cloud Computing

Xiangguo Liu, Yunpeng Luo, Anthony Goeckner, and 7 more authors

In 2023 60th ACM/IEEE Design Automation Conference (DAC), Jul 2023

Abs DOI Bib HTML

The rapid advancement of edge and cloud computing platforms, vehicular ad-hoc networks, and machine learning techniques have brought both opportunities and challenges for next-generation connected and automated vehicles (CAVs). On the one hand, these technologies can enable vehicles to leverage more computing power from edge and cloud servers and to share information with each other and surrounding infrastructures for better situation awareness and more intelligent decision making. On the other hand, the more distributed computing process and the wireless nature of V2X (vehicle-to-everything) communication expose vulnerabilities to various disturbances and attacks. In this paper, we discuss the security and safety challenges for edge- and cloud-enabled CAVs, particularly when they are under environment interferences, execution errors, and malicious attacks, and we will introduce our recent work and future directions in developing system-driven, end-to-end methodologies and tools to address these challenges and ensure system resiliency under uncertainties.
@inproceedings{liu_invited_2023, title = {Invited: {Waving} the {Double}-{Edged} {Sword}: {Building} {Resilient} {CAVs} with {Edge} and {Cloud} {Computing}}, copyright = {All rights reserved}, shorttitle = {Invited}, url = {https://ieeexplore.ieee.org/abstract/document/10247809}, doi = {10.1109/DAC56929.2023.10247809}, urldate = {2023-10-16}, booktitle = {2023 60th {ACM}/{IEEE} {Design} {Automation} {Conference} ({DAC})}, author = {Liu, Xiangguo and Luo, Yunpeng and Goeckner, Anthony and Chakraborty, Trishna and Jiao, Ruochen and Wang, Ningfei and Wang, Yixuan and Sato, Takami and Chen, Qi Alfred and Zhu, Qi}, month = jul, year = {2023}, pages = {1--4} }

ICML

Enforcing hard constraints with soft barriers: Safe reinforcement learning in unknown stochastic environments

Yixuan Wang, Simon Sinong Zhan, Ruochen Jiao, and 6 more authors

In International Conference on Machine Learning, 2023

Bib

@inproceedings{wang2023enforcing,
  title = {Enforcing hard constraints with soft barriers: Safe reinforcement learning in unknown stochastic environments},
  author = {Wang, Yixuan and Zhan, Simon Sinong and Jiao, Ruochen and Wang, Zhilu and Jin, Wanxin and Yang, Zhuoran and Wang, Zhaoran and Huang, Chao and Zhu, Qi},
  booktitle = {International Conference on Machine Learning},
  pages = {36593--36604},
  year = {2023},
  organization = {PMLR}
}

Empowering autonomous driving with large language models: A safety perspective

Yixuan Wang, Ruochen Jiao, Sinong Simon Zhan, and 5 more authors

arXiv preprint arXiv:2312.00812, 2023

Bib

@article{wang2023empowering,
  title = {Empowering autonomous driving with large language models: A safety perspective},
  author = {Wang, Yixuan and Jiao, Ruochen and Zhan, Sinong Simon and Lang, Chengtian and Huang, Chao and Wang, Zhaoran and Yang, Zhuoran and Zhu, Qi},
  journal = {arXiv preprint arXiv:2312.00812},
  year = {2023}
}

2022

Physics-Aware Safety-Assured Design of Hierarchical Neural Network based Planner

Xiangguo Liu, Chao Huang, Yixuan Wang, and 2 more authors

arXiv:2201.09140 [cs], Jan 2022

arXiv: 2201.09140

Abs Bib HTML

Neural networks have shown great promises in planning, control, and general decision making for learning-enabled cyber-physical systems (LE-CPSs), especially in improving performance under complex scenarios. However, it is very challenging to formally analyze the behavior of neural network based planners for ensuring system safety, which significantly impedes their applications in safety-critical domains such as autonomous driving. In this work, we propose a hierarchical neural network based planner that analyzes the underlying physical scenarios of the system and learns a system-level behavior planning scheme with multiple scenario-specific motion-planning strategies. We then develop an efficient verification method that incorporates overapproximation of the system state reachable set and novel partition and union techniques for formally ensuring system safety under our physics-aware planner. With theoretical analysis, we show that considering the different physical scenarios and building a hierarchical planner based on such analysis may improve system safety and verifiability. We also empirically demonstrate the effectiveness of our approach and its advantage over other baselines in practical case studies of unprotected left turn and highway merging, two common challenging safety-critical tasks in autonomous driving.
@article{liuPhysicsAwareSafetyAssuredDesign2022, title = {Physics-{Aware} {Safety}-{Assured} {Design} of {Hierarchical} {Neural} {Network} based {Planner}}, url = {http://arxiv.org/abs/2201.09140}, urldate = {2022-05-06}, journal = {arXiv:2201.09140 [cs]}, author = {Liu, Xiangguo and Huang, Chao and Wang, Yixuan and Zheng, Bowen and Zhu, Qi}, month = jan, year = {2022}, note = {arXiv: 2201.09140}, keywords = {Computer Science - Robotics} }
A Markov Decision Process framework to incorporate network-level data in motion planning for connected and automated vehicles

Xiangguo Liu, Neda Masoud, Qi Zhu, and 1 more author

Transportation Research Part C: Emerging Technologies, Mar 2022

Abs DOI Bib HTML

Autonomy and connectivity are expected to enhance safety and improve fuel efficiency in transportation systems. While connected vehicle-enabled technologies, such as coordinated cruise control, can improve vehicle motion planning by incorporating information beyond the line of sight of vehicles, their benefits are limited by the current short-sighted planning strategies that only utilize local information. In this paper, we propose a framework that devises vehicle trajectories by coupling a locally-optimal motion planner with a Markov decision process (MDP) model that can capture network-level information. Our proposed framework can guarantee safety while minimizing a trip’s generalized cost, which comprises of its fuel and time costs. To showcase the benefits of incorporating network-level data when devising vehicle trajectories, we conduct a comprehensive simulation study in three experimental settings, namely a circular track, a highway with on- and off-ramps, and a small urban network. The simulation results indicate that statistically significant efficiency can be obtained for the subject vehicle and its surrounding vehicles in different traffic states under all experimental settings. This paper serves as a proof-of-concept to showcase how connectivity and autonomy can be leveraged to incorporate network-level information into motion planning.
@article{liuMarkovDecisionProcess2022, title = {A {Markov} {Decision} {Process} framework to incorporate network-level data in motion planning for connected and automated vehicles}, volume = {136}, issn = {0968-090X}, url = {https://www.sciencedirect.com/science/article/pii/S0968090X21005325}, doi = {10.1016/j.trc.2021.103550}, language = {en}, urldate = {2022-05-06}, journal = {Transportation Research Part C: Emerging Technologies}, author = {Liu, Xiangguo and Masoud, Neda and Zhu, Qi and Khojandi, Anahita}, month = mar, year = {2022}, keywords = {Connected and automated vehicles, Trajectory planning}, pages = {103550} }
Design-while-verify: correct-by-construction control learning with verification in the loop

Yixuan Wang, Chao Huang, Zhaoran Wang, and 2 more authors

In Proceedings of the 59th ACM/IEEE Design Automation Conference, Aug 2022

Abs DOI Bib HTML

In the current control design of safety-critical cyber-physical systems, formal verification techniques are typically applied after the controller is designed to evaluate whether the required properties (e.g., safety) are satisfied. However, due to the increasing system complexity and the fundamental hardness of designing a controller with formal guarantees, such an open-loop process of design-then-verify often results in many iterations and fails to provide the necessary guarantees. In this paper, we propose a correct-by-construction control learning framework that integrates the verification into the control design process in a closed-loop manner, i.e., design-while-verify. Specifically, we leverage the verification results (computed reachable set of the system state) to construct feedback metrics for control learning, which measure how likely the current design of control parameters can meet the required reach-avoid property for safety and goal-reaching. We formulate an optimization problem based on such metrics for tuning the controller parameters, and develop an approximated gradient descent algorithm with a difference method to solve the optimization problem and learn the controller. The learned controller is formally guaranteed to meet the required reach-avoid property. By treating verifiability as a first-class objective and effectively leveraging the verification results during the control learning process, our approach can significantly improve the chance of finding a control design with formal property guarantees, demonstrated in a set of experiments that use model-based or neural network based controllers.
@inproceedings{wangDesignwhileverifyCorrectbyconstructionControl2022, address = {New York, NY, USA}, series = {{DAC} '22}, title = {Design-while-verify: correct-by-construction control learning with verification in the loop}, isbn = {978-1-4503-9142-9}, shorttitle = {Design-while-verify}, url = {https://dl.acm.org/doi/10.1145/3489517.3530556}, doi = {10.1145/3489517.3530556}, urldate = {2023-04-04}, booktitle = {Proceedings of the 59th {ACM}/{IEEE} {Design} {Automation} {Conference}}, publisher = {Association for Computing Machinery}, author = {Wang, Yixuan and Huang, Chao and Wang, Zhaoran and Wang, Zhilu and Zhu, Qi}, month = aug, year = {2022}, pages = {925--930} }
TAE: A Semi-supervised Controllable Behavior-aware Trajectory Generator and Predictor

Ruochen Jiao, Xiangguo Liu, Bowen Zheng, and 2 more authors

Nov 2022

arXiv:2203.01261 [cs]

Abs DOI Bib HTML

Trajectory generation and prediction are two interwoven tasks that play important roles in planner evaluation and decision making for intelligent vehicles. Most existing methods focus on one of the two and are optimized to directly output the final generated/predicted trajectories, which only contain limited information for critical scenario augmentation and safe planning. In this work, we propose a novel behavior-aware Trajectory Autoencoder (TAE) that explicitly models drivers’ behavior such as aggressiveness and intention in the latent space, using semi-supervised adversarial autoencoder and domain knowledge in transportation. Our model addresses trajectory generation and prediction in a unified architecture and benefits both tasks: the model can generate diverse, controllable and realistic trajectories to enhance planner optimization in safety-critical and long-tailed scenarios, and it can provide prediction of critical behavior in addition to the final trajectories for decision making. Experimental results demonstrate that our method achieves promising performance on both trajectory generation and prediction.
@misc{jiaoTAESemisupervisedControllable2022, title = {{TAE}: {A} {Semi}-supervised {Controllable} {Behavior}-aware {Trajectory} {Generator} and {Predictor}}, shorttitle = {{TAE}}, url = {http://arxiv.org/abs/2203.01261}, doi = {10.48550/arXiv.2203.01261}, urldate = {2023-04-04}, publisher = {arXiv}, author = {Jiao, Ruochen and Liu, Xiangguo and Zheng, Bowen and Liang, Dave and Zhu, Qi}, month = nov, year = {2022}, note = {arXiv:2203.01261 [cs]}, keywords = {Computer Science - Machine Learning, Computer Science - Robotics} }
Federated Class-Incremental Learning

Jiahua Dong, Lixu Wang, Zhen Fang, and 4 more authors

Mar 2022

arXiv:2203.11473 [cs]

Abs DOI Bib HTML

Federated learning (FL) has attracted growing attention via data-private collaborative training on decentralized clients. However, most existing methods unrealistically assume object classes of the overall framework are fixed over time. It makes the global model suffer from significant catastrophic forgetting on old classes in real-world scenarios, where local clients often collect new classes continuously and have very limited storage memory to store old classes. Moreover, new clients with unseen new classes may participate in the FL training, further aggravating the catastrophic forgetting of the global model. To address these challenges, we develop a novel Global-Local Forgetting Compensation (GLFC) model, to learn a global class incremental model for alleviating the catastrophic forgetting from both local and global perspectives. Specifically, to address local forgetting caused by class imbalance at the local clients, we design a class-aware gradient compensation loss and a class-semantic relation distillation loss to balance the forgetting of old classes and distill consistent inter-class relations across tasks. To tackle the global forgetting brought by the non-i.i.d class imbalance across clients, we propose a proxy server that selects the best old global model to assist the local relation distillation. Moreover, a prototype gradient-based communication mechanism is developed to protect privacy. Our model outperforms state-of-the-art methods by 4.4%-15.1% in terms of average accuracy on representative benchmark datasets.
@misc{dongFederatedClassIncrementalLearning2022, title = {Federated {Class}-{Incremental} {Learning}}, url = {http://arxiv.org/abs/2203.11473}, doi = {10.48550/arXiv.2203.11473}, urldate = {2023-04-04}, publisher = {arXiv}, author = {Dong, Jiahua and Wang, Lixu and Fang, Zhen and Sun, Gan and Xu, Shichao and Wang, Xiao and Zhu, Qi}, month = mar, year = {2022}, note = {arXiv:2203.11473 [cs]}, keywords = {Computer Science - Machine Learning} }
Non-Transferable Learning: A New Approach for Model Ownership Verification and Applicability Authorization

Lixu Wang, Shichao Xu, Ruiqi Xu, and 2 more authors

In , Jan 2022

Abs Bib HTML

As Artificial Intelligence as a Service gains popularity, protecting well-trained models as intellectual property is becoming increasingly important. There are two common types of protection methods: ownership verification and usage authorization. In this paper, we propose Non-Transferable Learning (NTL), a novel approach that captures the exclusive data representation in the learned model and restricts the model generalization ability to certain domains. This approach provides effective solutions to both model verification and authorization. Specifically: 1) For ownership verification, watermarking techniques are commonly used but are often vulnerable to sophisticated watermark removal methods. By comparison, our NTL-based ownership verification provides robust resistance to state-of-the-art watermark removal methods, as shown in extensive experiments with 6 removal approaches over the digits, CIFAR10 & STL10, and VisDA datasets. 2) For usage authorization, prior solutions focus on authorizing specific users to access the model, but authorized users can still apply the model to any data without restriction. Our NTL-based authorization approach instead provides data-centric protection, which we call applicability authorization, by significantly degrading the performance of the model on unauthorized data. Its effectiveness is also shown through experiments on aforementioned datasets.
@inproceedings{wangNonTransferableLearningNew2022, title = {Non-{Transferable} {Learning}: {A} {New} {Approach} for {Model} {Ownership} {Verification} and {Applicability} {Authorization}}, shorttitle = {Non-{Transferable} {Learning}}, url = {https://openreview.net/forum?id=tYRrOdSnVUy}, language = {en}, urldate = {2023-04-04}, author = {Wang, Lixu and Xu, Shichao and Xu, Ruiqi and Wang, Xiao and Zhu, Qi}, month = jan, year = {2022} }
Speech Disfluency Detection with Contextual Representation and Data Distillation

Payal Mohapatra, Akash Pandey, Bashima Islam, and 1 more author

In Proceedings of the 1st ACM International Workshop on Intelligent Acoustic Systems and Applications, Jun 2022

Abs DOI Bib HTML

Stuttering affects almost 1}% of the world’s population. It has a deep sociological impact and hinders the people who stutter from taking advantage of voice-assisted services. Automatic stutter detection based on deep learning can help voice assistants to adapt themselves to atypical speech. However, disfluency data is very limited and expensive to generate. We propose a set of preprocessing techniques: (1) using data with high inter-annotator agreement, (2) balancing different classes, and (3) using contextual embeddings from a pretrained network. We then design a disfluency classification network (DisfluencyNet) for automated speech disfluency detection that takes these contextual embeddings as an input. We empirically demonstrate high performance using only a quarter of the data for training. We conduct experiments with different training data size, evaluate the model trained on the lowest amount of training data with SEP-28k baseline results, and evaluate the same model on the FluencyBank dataset baseline results. We observe that, even by using a quarter of the original size of the dataset, our F1 score is greater than 0.7 for all types of disfluencies except one,}textit{ blocks}. Previous works also reported lower performance with }textit{blocks} type of disfluency owing to its large diversity amongst speakers and events. Overall, with our approach using only a few minutes of data, we can train a robust network that outperforms the baseline results for all disfluencies by at least 5}%. Such a result is important to stress the fact that we can now reduce the required amount of training data and are able to improve the quality of the dataset by appointing more than two annotators for labeling speech disfluency within a constrained labeling budget.
@inproceedings{mohapatraSpeechDisfluencyDetection2022, address = {New York, NY, USA}, series = {{IASA} '22}, title = {Speech {Disfluency} {Detection} with {Contextual} {Representation} and {Data} {Distillation}}, isbn = {978-1-4503-9403-1}, url = {https://dl.acm.org/doi/10.1145/3539490.3539601}, doi = {10.1145/3539490.3539601}, urldate = {2023-04-04}, booktitle = {Proceedings of the 1st {ACM} {International} {Workshop} on {Intelligent} {Acoustic} {Systems} and {Applications}}, publisher = {Association for Computing Machinery}, author = {Mohapatra, Payal and Pandey, Akash and Islam, Bashima and Zhu, Qi}, month = jun, year = {2022}, keywords = {Deep Learning, Contextual Representation, Neural Networks, Speech Disfluency}, pages = {19--24} }

2021

Cross-Layer Design of Automotive Systems

Zhilu Wang, Hengyi Liang, Chao Huang, and 1 more author

IEEE Design Test, Oct 2021

Conference Name: IEEE Design Test

Abs DOI Bib

This article presents multiple cross-layer methods to ensure security as well as functional correctness of automotive systems. These approaches tie together multiple abstraction layers, going all the way from vehicular networks responsible for vehicle-to-vehicle and vehicle-to-infrastructure communication, to in-vehicle hardware/software architectures.

@article{wangCrossLayerDesignAutomotive2021,
  title = {Cross-{Layer} {Design} of {Automotive} {Systems}},
  volume = {38},
  issn = {2168-2364},
  doi = {10.1109/MDAT.2020.3037561},
  number = {5},
  journal = {IEEE Design Test},
  author = {Wang, Zhilu and Liang, Hengyi and Huang, Chao and Zhu, Qi},
  month = oct,
  year = {2021},
  note = {Conference Name: IEEE Design Test},
  keywords = {Security, Computer architecture, Connected vehicles, Automotive engineering, Vehicular ad hoc networks, Delays, Automotive system, Cross layer design, Cross-layer design, Cyber-physical system, Cyper-physical systems, Weakly-hard paradigm},
  pages = {8--16}
}

Safety-Assured Design and Adaptation of Learning-Enabled Autonomous Systems

Qi Zhu, Chao Huang, Ruochen Jiao, and 6 more authors

In Proceedings of the 26th Asia and South Pacific Design Automation Conference, Jan 2021

Abs DOI Bib HTML

Future autonomous systems will employ sophisticated machine learning techniques for the sensing and perception of the surroundings and the making corresponding decisions for planning, control, and other actions. They often operate in highly dynamic, uncertain and challenging environment, and need to meet stringent timing, resource, and mission requirements. In particular, it is critical and yet very challenging to ensure the safety of these autonomous systems, given the uncertainties of the system inputs, the constant disturbances on the system operations, and the lack of analyzability for many machine learning methods (particularly those based on neural networks). In this paper, we will discuss some of these challenges, and present our work in developing automated, quantitative, and formalized methods and tools for ensuring the safety of autonomous systems in their design and during their runtime adaptation. We argue that it is essential to take a holistic approach in addressing system safety and other safety-related properties, vertically across the functional, software, and hardware layers, and horizontally across the autonomy pipeline of sensing, perception, planning, and control modules. This approach could be further extended from a single autonomous system to a multi-agent system where multiple autonomous agents perform tasks in a collaborative manner. We will use connected and autonomous vehicles (CAVs) as the main application domain to illustrate the importance of such holistic approach and show our initial efforts in this direction.
@inproceedings{zhuSafetyAssuredDesignAdaptation2021, address = {New York, NY, USA}, series = {{ASPDAC} '21}, title = {Safety-{Assured} {Design} and {Adaptation} of {Learning}-{Enabled} {Autonomous} {Systems}}, isbn = {978-1-4503-7999-1}, url = {https://doi.org/10.1145/3394885.3431623}, doi = {10.1145/3394885.3431623}, urldate = {2022-05-05}, booktitle = {Proceedings of the 26th {Asia} and {South} {Pacific} {Design} {Automation} {Conference}}, publisher = {Association for Computing Machinery}, author = {Zhu, Qi and Huang, Chao and Jiao, Ruochen and Lan, Shuyue and Liang, Hengyi and Liu, Xiangguo and Wang, Yixuan and Wang, Zhilu and Xu, Shichao}, month = jan, year = {2021}, pages = {753--760} }
DATE
Adaptive Learning Based Building Load Prediction for Microgrid Economic Dispatch

Rumia Masburah, Rajib Lochan Jana, Ainuddin Khan, and 4 more authors

In 2021 Design, Automation Test in Europe Conference Exhibition (DATE), Feb 2021

ISSN: 1558-1101

Abs DOI Bib

Given that building loads consume roughly 40% of the energy produced in developed countries, smart buildings with local renewable resources offer a viable alternative towards achieving a greener future. Building temperature control strategies typically employ detailed physical models which require a significant amount of time, information and finesse. Even then, due to unknown building parameters and related inaccuracies, future power demands by the building loads are difficult to estimate. This creates unique challenges in the domain of microgrid economic power dispatch for satisfying building power demands through efficient control and scheduling of renewable and non-renewable local resources in conjunction with supply from the main grid. In this work, we estimate the real-time uncertainties in building loads using Gaussian Process (GP) learning and establish the effectiveness of run time model correction in the context of microgrid economic dispatch.
@inproceedings{masburahAdaptiveLearningBased2021, title = {Adaptive {Learning} {Based} {Building} {Load} {Prediction} for {Microgrid} {Economic} {Dispatch}}, doi = {10.23919/DATE51398.2021.9474041}, booktitle = {2021 {Design}, {Automation} {Test} in {Europe} {Conference} {Exhibition} ({DATE})}, author = {Masburah, Rumia and Jana, Rajib Lochan and Khan, Ainuddin and Xu, Shichao and Lan, Shuyue and Dey, Soumyajit and Zhu, Qi}, month = feb, year = {2021}, note = {ISSN: 1558-1101}, keywords = {Adaptation models, Buildings, Uncertainty, Biological system modeling, Building thermal model, Deep Reinforcement Learning, Economic Dispatch, Economics, Gaussian Process Learning, Microgrids, Power demand, Predictive Control}, pages = {72--75} }
DATE
Bounding Perception Neural Network Uncertainty for Safe Control of Autonomous Systems

Zhilu Wang, Chao Huang, Yixuan Wang, and 3 more authors

In 2021 Design, Automation Test in Europe Conference Exhibition (DATE), Feb 2021

ISSN: 1558-1101

Abs DOI Bib

Future autonomous systems will rely on advanced sensors and deep neural networks for perceiving the environment, and then utilize the perceived information for system planning, control, adaptation, and general decision making. However, due to the inherent uncertainties from the dynamic environment and the lack of methodologies for predicting neural network behavior, the perception modules in autonomous systems often could not provide deterministic guarantees and may sometimes lead the system into unsafe states (e.g., as evident by a number of high-profile accidents with experimental autonomous vehicles). This has significantly impeded the broader application of machine learning techniques, particularly those based on deep neural networks, in safety-critical systems. In this paper, we will discuss these challenges, define open research problems, and introduce our recent work in developing formal methods for quantitatively bounding the output uncertainty of perception neural networks with respect to input perturbations, and leveraging such bounds to formally ensure the safety of system control. Unlike most existing works that only focus on either the perception module or the control module, our approach provides a holistic end-to-end framework that bounds the perception uncertainty and addresses its impact on control.
@inproceedings{wangBoundingPerceptionNeural2021, title = {Bounding {Perception} {Neural} {Network} {Uncertainty} for {Safe} {Control} of {Autonomous} {Systems}}, doi = {10.23919/DATE51398.2021.9474204}, booktitle = {2021 {Design}, {Automation} {Test} in {Europe} {Conference} {Exhibition} ({DATE})}, author = {Wang, Zhilu and Huang, Chao and Wang, Yixuan and Hobbs, Clara and Chakraborty, Samarjit and Zhu, Qi}, month = feb, year = {2021}, note = {ISSN: 1558-1101}, keywords = {Neural networks, Safety, Control systems, Uncertainty, Perturbation methods, Runtime, Sensor systems}, pages = {1745--1750} }
RTAS
Brief Industry Paper: An Infrastructure-Aided High Definition Map Data Provisioning Service for Autonomous Driving

Jinliang Xie, Jie Tang, Yanzhi Wang, and 2 more authors

In 2021 IEEE 27th Real-Time and Embedded Technology and Applications Symposium (RTAS), May 2021

ISSN: 2642-7346

Abs DOI Bib

As a fundamental component in the autonomous driving technology stack, High Definition Maps (HD map) provide high-precision descriptions of the environment. It enables extremely accurate perception and localization while improving the efficiency of path planning. However, the HD map’s extremely large data volume poses great challenges for the real-time and safety requirements of autonomous driving. Based on our real-world deployment experiences, we first demonstrate how the existing data transmission mechanism is weak in supporting HD map services. To address this problem, we propose an HD map data service mechanism on top of Vehicle-to-Infrastructure (V2I) data transmission under a tight time and energy budget. By this mechanism, the selected road side unit (RSU) nodes cooperate on map provisioning tasks and transmit HD map data proportionately. Furthermore, we model the real-time map data service into a partial knapsack problem and develop a greedy data transmission algorithm. Experimental results confirm that the proposed mechanism can ensure the real-time HD map data service meanwhile meeting the energy limits.
@inproceedings{xieBriefIndustryPaper2021, title = {Brief {Industry} {Paper}: {An} {Infrastructure}-{Aided} {High} {Definition} {Map} {Data} {Provisioning} {Service} for {Autonomous} {Driving}}, shorttitle = {Brief {Industry} {Paper}}, doi = {10.1109/RTAS52030.2021.00042}, booktitle = {2021 {IEEE} 27th {Real}-{Time} and {Embedded} {Technology} and {Applications} {Symposium} ({RTAS})}, author = {Xie, Jinliang and Tang, Jie and Wang, Yanzhi and Zhu, Qi and Liu, Shaoshan}, month = may, year = {2021}, note = {ISSN: 2642-7346}, keywords = {Safety, Autonomous driving, Energy Efficiency, HD Maps, Location awareness, Map Data distribution, Map Data provisioning, Meetings, Path planning, Real-time systems, Road side unit, Vehicle-to-infrastructure}, pages = {421--424} }
Cross-Layer Adaptation with Safety-Assured Proactive Task Job Skipping

Zhilu Wang, Chao Huang, Hyoseung Kim, and 2 more authors

ACM Transactions on Embedded Computing Systems, Sep 2021

Abs DOI Bib HTML

During the operation of many real-time safety-critical systems, there are often strong needs for adapting to a dynamic environment or evolving mission objectives, e.g., increasing sampling and control frequencies of some functions to improve their performance under certain situations. However, a system’s ability to adapt is often limited by tight resource constraints and rigid periodic execution requirements. In this work, we present a cross-layer approach to improve system adaptability by allowing proactive skipping of task executions, so that the resources can be either saved directly or re-allocated to other tasks for their performance improvement. Our approach includes three novel elements: (1) formal methods for deriving the feasible skipping choices of control tasks with safety guarantees at the functional layer, (2) a schedulability analysis method for assessing system feasibility at the architectural layer under allowed task job skippings, and (3) a runtime adaptation algorithm that efficiently explores job skipping choices and task priorities for meeting system adaptation requirements while ensuring system safety and timing correctness. Experiments demonstrate the effectiveness of our approach in meeting system adaptation needs.
@article{wangCrossLayerAdaptationSafetyAssured2021, title = {Cross-{Layer} {Adaptation} with {Safety}-{Assured} {Proactive} {Task} {Job} {Skipping}}, volume = {20}, issn = {1539-9087}, url = {https://doi.org/10.1145/3477031}, doi = {10.1145/3477031}, number = {5s}, urldate = {2022-05-06}, journal = {ACM Transactions on Embedded Computing Systems}, author = {Wang, Zhilu and Huang, Chao and Kim, Hyoseung and Li, Wenchao and Zhu, Qi}, month = sep, year = {2021}, keywords = {adaptation, Cross-layer, safety, weakly hard}, pages = {100:1--100:25} }

Weak Adaptation Learning: Addressing Cross-Domain Data Insufficiency With Weak Annotator

Shichao Xu, Lixu Wang, Yixuan Wang, and 1 more author

In , 2021

Bib HTML

@inproceedings{xuWeakAdaptationLearning2021,
  title = {Weak {Adaptation} {Learning}: {Addressing} {Cross}-{Domain} {Data} {Insufficiency} {With} {Weak} {Annotator}},
  shorttitle = {Weak {Adaptation} {Learning}},
  url = {https://openaccess.thecvf.com/content/ICCV2021/html/Xu_Weak_Adaptation_Learning_Addressing_Cross-Domain_Data_Insufficiency_With_Weak_Annotator_ICCV_2021_paper.html},
  language = {en},
  urldate = {2022-05-06},
  author = {Xu, Shichao and Wang, Lixu and Wang, Yixuan and Zhu, Qi},
  year = {2021},
  pages = {8917--8926}
}

Learning-based framework for sensor fault-tolerant building HVAC control with model-assisted learning

Shichao Xu, Yangyang Fu, Yixuan Wang, and 2 more authors

In Proceedings of the 8th ACM International Conference on Systems for Energy-Efficient Buildings, Cities, and Transportation, Nov 2021

Abs DOI Bib HTML

As people spend up to 87% of their time indoors, intelligent Heating, Ventilation, and Air Conditioning (HVAC) systems in buildings are essential for maintaining occupant comfort and reducing energy consumption. These HVAC systems in smart buildings rely ’on real-time sensor readings, which in practice often suffer from various faults and could also be vulnerable to malicious attacks. Such faulty sensor inputs may lead to the violation of indoor environment requirements (e.g., temperature, humidity, etc.) and the increase of energy consumption. While many model-based approaches have been proposed in the literature for building HVAC control, it is costly to develop accurate physical models for ensuring their performance and even more challenging to address the impact of sensor faults. In this work, we present a novel learning-based framework for sensor fault-tolerant HVAC control, which includes three deep learning based components for 1) generating temperature proposals with the consideration of possible sensor faults, 2) selecting one of the proposals based on the assessment of their accuracy, and 3) applying reinforcement learning with the selected temperature proposal. Moreover, to address the challenge of training data insufficiency in building-related tasks, we propose a model-assisted learning method leveraging an abstract model of building physical dynamics. Through extensive experiments, we demonstrate that the proposed fault-tolerant HVAC control framework can significantly reduce building temperature violations under a variety of sensor fault patterns while maintaining energy efficiency.
@inproceedings{xuLearningbasedFrameworkSensor2021, address = {New York, NY, USA}, series = {{BuildSys} '21}, title = {Learning-based framework for sensor fault-tolerant building {HVAC} control with model-assisted learning}, isbn = {978-1-4503-9114-6}, url = {https://doi.org/10.1145/3486611.3486644}, doi = {10.1145/3486611.3486644}, urldate = {2022-05-05}, booktitle = {Proceedings of the 8th {ACM} {International} {Conference} on {Systems} for {Energy}-{Efficient} {Buildings}, {Cities}, and {Transportation}}, publisher = {Association for Computing Machinery}, author = {Xu, Shichao and Fu, Yangyang and Wang, Yixuan and O'Neill, Zheng and Zhu, Qi}, month = nov, year = {2021}, keywords = {deep learning, HVAC control, sensor fault-tolerant}, pages = {1--10} }
DAC
Invited: Towards Fully Intelligent Transportation through Infrastructure-Vehicle Cooperative Autonomous Driving: Challenges and Opportunities

Shaoshan Liu, Bo Yu, Jie Tang, and 1 more author

In 2021 58th ACM/IEEE Design Automation Conference (DAC), Dec 2021

ISSN: 0738-100X

Abs DOI Bib

The infrastructure-vehicle cooperative autonomous driving approach relies on the cooperation between intelligent roads and intelligent vehicles. This approach is not only safer but also more economical compared to the traditional on-vehicle-only autonomous driving. In this paper, we introduce the real-world deployment experiences of infrastructure-vehicle cooperative autonomous driving by PerceptIn, where a three-stage development roadmap is taken: infrastructure-augmented autonomous driving (IAAD), infrastructure-guided autonomous driving (IGAD), and infrastructure-planned autonomous driving (IPAD). We then discuss the future research challenges and opportunities for such approach.
@inproceedings{liuInvitedFullyIntelligent2021, title = {Invited: {Towards} {Fully} {Intelligent} {Transportation} through {Infrastructure}-{Vehicle} {Cooperative} {Autonomous} {Driving}: {Challenges} and {Opportunities}}, shorttitle = {Invited}, doi = {10.1109/DAC18074.2021.9586317}, booktitle = {2021 58th {ACM}/{IEEE} {Design} {Automation} {Conference} ({DAC})}, author = {Liu, Shaoshan and Yu, Bo and Tang, Jie and Zhu, Qi}, month = dec, year = {2021}, note = {ISSN: 0738-100X}, keywords = {Transportation, Roads, Autonomous vehicles, Design automation, Tablet computers}, pages = {1323--1326} }
DAC
Cocktail: Learn a Better Neural Network Controller from Multiple Experts via Adaptive Mixing and Robust Distillation

Yixuan Wang, Chao Huang, Zhilu Wang, and 3 more authors

In 2021 58th ACM/IEEE Design Automation Conference (DAC), Dec 2021

ISSN: 0738-100X

Abs DOI Bib

Neural networks are being increasingly applied to control and decision making for learning-enabled cyber-physical systems (LE-CPSs). They have shown promising performance without requiring the development of complex physical models; however, their adoption is significantly hindered by the concerns on their safety, robustness, and efficiency. In this work, we propose COCKTAIL, a novel design framework that automatically learns a neural network based controller from multiple existing control methods (experts) that could be either model-based or neural network based. In particular, COCKTAIL first performs reinforcement learning to learn an optimal system-level adaptive mixing strategy that incorporates the underlying experts with dynamically-assigned weights, and then conducts a teacher-student distillation with probabilistic adversarial training and regularization to synthesize a student neural network controller with improved control robustness (measured by a safe control rate metric with respect to adversarial attacks or measurement noises), control energy efficiency, and verifiability (measured by the computation time for verification). Experiments on three non-linear systems demonstrate significant advantages of our approach on these properties over various baseline methods.
@inproceedings{wangCocktailLearnBetter2021, title = {Cocktail: {Learn} a {Better} {Neural} {Network} {Controller} from {Multiple} {Experts} via {Adaptive} {Mixing} and {Robust} {Distillation}}, shorttitle = {Cocktail}, doi = {10.1109/DAC18074.2021.9586148}, booktitle = {2021 58th {ACM}/{IEEE} {Design} {Automation} {Conference} ({DAC})}, author = {Wang, Yixuan and Huang, Chao and Wang, Zhilu and Xu, Shichao and Wang, Zhaoran and Zhu, Qi}, month = dec, year = {2021}, note = {ISSN: 0738-100X}, keywords = {Reinforcement learning, Robustness, Adaptive systems, Energy measurement, Neural networks, Time measurement, Weight measurement}, pages = {397--402} }
Deep Reinforcement Learning for Joint Datacenter and HVAC Load Control in Distributed Mixed-Use Buildings

Tianshu Wei, Shaolei Ren, and Qi Zhu

IEEE Transactions on Sustainable Computing, Jul 2021

Conference Name: IEEE Transactions on Sustainable Computing

Abs DOI Bib

The majority of today’s power-hungry datacenters are physically co-located with office rooms in mixed-use buildings (MUBs). The heating, ventilation, and air conditioning (HVAC) system within each MUB is often shared or partially-shared between datacenter rooms and office zones, for removing the heat generated by computing equipment and maintaining desired room temperature for building tenants. To effectively reduce the total energy cost of MUBs, it is important to leverage the scheduling flexibility in both the HVAC system and the datacenter workload. In this work, we formulate both HVAC control and datacenter workload scheduling as a Markov decision process (MDP), and propose a deep reinforcement learning (DRL) based algorithm for minimizing the total energy cost while maintaining desired room temperature and meeting datacenter workload deadline constraints. Moreover, we also develop a heuristic DRL-based algorithm to enable interactive workload allocation among geographically distributed MUBs for further energy reduction. The experiment results demonstrate that our regular DRL-based algorithm can achieve up to 26.9 percent cost reduction for a single MUB, when compared with a baseline strategy. Our heuristic DRL-based algorithm can reduce the total energy cost by an additional 5.5 percent, when intelligently allocating interactive workload for multiple geographically distributed MUBs.
@article{weiDeepReinforcementLearning2021, title = {Deep {Reinforcement} {Learning} for {Joint} {Datacenter} and {HVAC} {Load} {Control} in {Distributed} {Mixed}-{Use} {Buildings}}, volume = {6}, issn = {2377-3782}, doi = {10.1109/TSUSC.2019.2910533}, number = {3}, journal = {IEEE Transactions on Sustainable Computing}, author = {Wei, Tianshu and Ren, Shaolei and Zhu, Qi}, month = jul, year = {2021}, note = {Conference Name: IEEE Transactions on Sustainable Computing}, keywords = {Heuristic algorithms, Deep reinforcement learning, Buildings, Servers, Energy consumption, Processor scheduling, Load modeling, datacenter, geographically distributed, HVAC, mixed-use buildings}, pages = {370--384} }

WIP: End-to-End Analysis of Adversarial Attacks to Automated Lane Centering Systems

Hengyi Liang, Ruochen Jiao, Takami Sato, and 3 more authors

Workshop on Automotive and Autonomous Vehicle Security (AutoSec’21), Feb 2021

Bib HTML

@article{liangWIPEndtoEndAnalysis2021,
  title = {{WIP}: {End}-to-{End} {Analysis} of {Adversarial} {Attacks} to {Automated} {Lane} {Centering} {Systems}},
  shorttitle = {{WIP}},
  url = {https://par.nsf.gov/biblio/10289738-wip-end-end-analysis-adversarial-attacks-automated-lane-centering-systems},
  language = {en},
  urldate = {2022-05-06},
  journal = {Workshop on Automotive and Autonomous Vehicle Security (AutoSec'21)},
  author = {Liang, Hengyi and Jiao, Ruochen and Sato, Takami and Shen, Junjie and Chen, Qi Alfred and Zhu, Qi},
  month = feb,
  year = {2021}
}

IV
Credibility Enhanced Temporal Graph Convolutional Network Based Sybil Attack Detection On Edge Computing Servers

Baiting Luo, Xiangguo Liu, and Qi Zhu

In 2021 IEEE Intelligent Vehicles Symposium (IV), Jul 2021

Abs DOI Bib

The emerging vehicular edge computing (VEC) technology has the potential to bring revolutionary development to vehicular ad hoc network (VANET). However, the edge computing servers (ECSs) are subjected to a variety of security threats. One of the most dangerous types of security attacks is the Sybil attack, which can create fabricated virtual vehicles (called Sybil vehicles) to significantly overload ECSs’ limited computation resources and thus disrupt legitimate vehicles’ edge computing applications. In this paper, we present a novel Sybil attack detection system on ECSs that is based on the design of a credibility enhanced temporal graph convolutional network. Our approach can identify the malicious vehicles in a dynamic traffic environment while preserving the legitimate vehicles’ privacy, particularly their local position information. We evaluate our proposed approach in the SUMO simulator. The results demonstrate that our proposed detection system can accurately identify most Sybil vehicles while maintaining a low error rate.
@inproceedings{luoCredibilityEnhancedTemporal2021, title = {Credibility {Enhanced} {Temporal} {Graph} {Convolutional} {Network} {Based} {Sybil} {Attack} {Detection} {On} {Edge} {Computing} {Servers}}, doi = {10.1109/IV48863.2021.9575361}, booktitle = {2021 {IEEE} {Intelligent} {Vehicles} {Symposium} ({IV})}, author = {Luo, Baiting and Liu, Xiangguo and Zhu, Qi}, month = jul, year = {2021}, keywords = {Privacy, Simulation, Vehicular ad hoc networks, Servers, Real-time systems, Error analysis, System performance}, pages = {524--531} }
IV
Securing Connected Vehicle Applications with an Efficient Dual Cyber- Physical Blockchain Framework

Xiangguo Liu, Baiting Luo, Ahmed Abdo, and 2 more authors

In 2021 IEEE Intelligent Vehicles Symposium (IV), Jul 2021

Abs DOI Bib

While connected vehicle (CV) applications have the potential to revolutionize traditional transportation system, cyber and physical attacks on them may lead to disastrous consequences. In this work, we propose an efficient dual cyber-physical blockchain framework to build trust and secure communication for CV applications. Our approach incorporates blockchain technology and physical sensing capabilities of vehicles to quickly react to attacks in a large-scale vehicular network, with low resource overhead. We explore the application of our framework to three CV applications, i.e., highway merging, intelligent intersection management, and traffic network with route choices. Simulation results demonstrate the effectiveness of our blockchain-based framework in defending against spoofing attacks, bad mouthing attacks, and Sybil and voting attacks. We also provide analysis to show the timing and resource efficiency of our framework.
@inproceedings{liuSecuringConnectedVehicle2021, title = {Securing {Connected} {Vehicle} {Applications} with an {Efficient} {Dual} {Cyber}- {Physical} {Blockchain} {Framework}}, doi = {10.1109/IV48863.2021.9575869}, booktitle = {2021 {IEEE} {Intelligent} {Vehicles} {Symposium} ({IV})}, author = {Liu, Xiangguo and Luo, Baiting and Abdo, Ahmed and Abu-Ghazaleh, Nael and Zhu, Qi}, month = jul, year = {2021}, keywords = {Buildings, Connected vehicles, Intelligent vehicles, Merging, Road transportation, Simulation, Transportation}, pages = {393--400} }
IV
End-to-end Uncertainty-based Mitigation of Adversarial Attacks to Automated Lane Centering

Ruochen Jiao, Hengyi Liang, Takami Sato, and 3 more authors

In 2021 IEEE Intelligent Vehicles Symposium (IV), Jul 2021

Abs DOI Bib

In the development of advanced driver-assistance systems (ADAS) and autonomous vehicles, machine learning techniques that are based on deep neural networks (DNNs) have been widely used for vehicle perception. These techniques offer significant improvement on average perception accuracy over traditional methods, however have been shown to be susceptible to adversarial attacks, where small perturbations in the input may cause significant errors in the perception results and lead to system failure. Most prior works addressing such adversarial attacks focus only on the sensing and perception modules. In this work, we propose an end-to-end approach that addresses the impact of adversarial attacks throughout perception, planning, and control modules. In particular, we choose a target ADAS application, the automated lane centering system in OpenPilot, quantify the perception uncertainty under adversarial attacks, and design a robust planning and control module accordingly based on the uncertainty analysis. We evaluate our proposed approach using both public dataset and production-grade autonomous driving simulator. The experiment results demonstrate that our approach can effectively mitigate the impact of adversarial attack and can achieve 55% 90% improvement over the original OpenPilot.
@inproceedings{jiaoEndtoendUncertaintybasedMitigation2021, title = {End-to-end {Uncertainty}-based {Mitigation} of {Adversarial} {Attacks} to {Automated} {Lane} {Centering}}, doi = {10.1109/IV48863.2021.9575549}, booktitle = {2021 {IEEE} {Intelligent} {Vehicles} {Symposium} ({IV})}, author = {Jiao, Ruochen and Liang, Hengyi and Sato, Takami and Shen, Junjie and Chen, Qi Alfred and Zhu, Qi}, month = jul, year = {2021}, keywords = {Adaptation models, Planning, Deep learning, Estimation, Data models, Uncertainty, Perturbation methods}, pages = {266--273} }

2020

ICCAD
Energy-Efficient Control Adaptation with Safety Guarantees for Learning-Enabled Cyber-Physical Systems

Yixuan Wang, Chao Huang, and Qi Zhu

In 2020 IEEE/ACM International Conference On Computer Aided Design (ICCAD), Nov 2020

ISSN: 1558-2434

Abs Bib

Neural networks have been increasingly applied to control in learning-enabled cyber-physical systems (LE-CPSs) and demonstrated great promises in improving system performance and efficiency, as well as reducing the need for complex physical models. However, the lack of safety guarantees for such neural network based controllers has significantly impeded their adoption in safety-critical CPSs. In this work, we propose a controller adaptation approach that automatically switches among multiple controllers, including neural network controllers, to guarantee system safety and improve energy efficiency. Our approach includes two key components based on formal methods and machine learning. First, we approximate each controller with a Bernstein-polynomial based hybrid system model under bounded disturbance, and compute a safe invariant set for each controller based on its corresponding hybrid system. Intuitively, the invariant set of a controller defines the state space where the system can always remain safe under its control. The union of the controllers’ invariants sets then define a safe adaptation space that is larger than (or equal to) that of each controller. Second, we develop a deep reinforcement learning method to learn a controller switching strategy for reducing the control/actuation energy cost, while with the help of a safety guard rule, ensuring that the system stays within the safe space. Experiments on a linear adaptive cruise control system and a non-linear Van der Pol’s oscillator demonstrate the effectiveness of our approach on energy saving and safety enhancement.
@inproceedings{wangEnergyEfficientControlAdaptation2020, title = {Energy-{Efficient} {Control} {Adaptation} with {Safety} {Guarantees} for {Learning}-{Enabled} {Cyber}-{Physical} {Systems}}, booktitle = {2020 {IEEE}/{ACM} {International} {Conference} {On} {Computer} {Aided} {Design} ({ICCAD})}, author = {Wang, Yixuan and Huang, Chao and Zhu, Qi}, month = nov, year = {2020}, note = {ISSN: 1558-2434}, keywords = {Aerospace electronics, adaptation, Neural networks, Safety, Control systems, Energy consumption, Energy efficiency, invariant, LE-CPS, neural network, Oscillators, Safety guarantees}, pages = {1--9} }
Trajectory Planning for Connected and Automated Vehicles: Cruising, Lane Changing, and Platooning

Xiangguo Liu, Guangchen Zhao, Neda Masoud, and 1 more author

arXiv:2001.08620 [cs], Jan 2020

arXiv: 2001.08620

Abs Bib HTML

Autonomy and connectivity are considered among the most promising technologies to improve safety, mobility, fuel and time consumption in transportation systems. Some of the fuel efficiency benefits of connected and automated vehicles (CAVs) can be realized through platooning. A platoon is a virtual train of CAVs that travel together following the platoon head, with small gaps between them. Vehicles may also reduce travel time by lane changing. In this paper, we devise an optimal control-based trajectory planning model that can provide safe and efficient trajectories for the subject vehicle and can incorporate platooning and lane changing. We embed this trajectory planning model in a simulation framework to quantify its efficiency benefits as it relates to fuel consumption and travel time, in a dynamic traffic stream. Furthermore, we perform extensive numerical experiments to investigate whether, and the circumstances under which, the vehicles in upstream of the subject vehicle may also experience second-hand fuel efficiency benefits.
@article{liuTrajectoryPlanningConnected2020, title = {Trajectory {Planning} for {Connected} and {Automated} {Vehicles}: {Cruising}, {Lane} {Changing}, and {Platooning}}, shorttitle = {Trajectory {Planning} for {Connected} and {Automated} {Vehicles}}, url = {http://arxiv.org/abs/2001.08620}, urldate = {2022-05-06}, journal = {arXiv:2001.08620 [cs]}, author = {Liu, Xiangguo and Zhao, Guangchen and Masoud, Neda and Zhu, Qi}, month = jan, year = {2020}, note = {arXiv: 2001.08620}, keywords = {Computer Science - Robotics} }

MaskPlus: Improving Mask Generation for Instance Segmentation

Shichao Xu, Shuyue Lan, and Zhu Qi

In , 2020

Bib HTML

@inproceedings{xuMaskPlusImprovingMask2020,
  title = {{MaskPlus}: {Improving} {Mask} {Generation} for {Instance} {Segmentation}},
  shorttitle = {{MaskPlus}},
  url = {https://openaccess.thecvf.com/content_WACV_2020/html/Xu_MaskPlus_Improving_Mask_Generation_for_Instance_Segmentation_WACV_2020_paper.html},
  urldate = {2022-05-06},
  author = {Xu, Shichao and Lan, Shuyue and Qi, Zhu},
  year = {2020},
  pages = {2030--2038}
}

IV
Impact of Sharing Driving Attitude Information: A Quantitative Study on Lane Changing

Xiangguo Liu, Neda Masoud, and Qi Zhu

In 2020 IEEE Intelligent Vehicles Symposium (IV), Oct 2020

ISSN: 2642-7214

Abs DOI Bib

Autonomous vehicles (AVs) are expected to be an integral part of the next generation of transportation systems, where they will share the transportation network with human-driven vehicles during the transition period. In this work, we model the interactions between vehicles (two AVs or an AV and a human-driven vehicle) in a lane changing process by leveraging the Stackelberg game. We explicitly model driving attitudes for both vehicles involved in lane changing. We design five cases, in which the two vehicles have different levels of knowledge, and make different assumptions, about the driving attitude of the rival. We conduct theoretical analysis and simulations for different cases in two lane changing scenarios, namely changing lanes from a higher-speed lane to a lower-speed lane, and from a lower-speed lane to a higher-speed lane. We use four metrics (fuel consumption, discomfort, minimum distance gap and lane change success rate) to investigate how the performance of a single vehicle and that of the system will be influenced by the level of information sharing, and whether a vehicle trajectory optimized based on selfish criteria can provide system-level benefits.
@inproceedings{liuImpactSharingDriving2020, title = {Impact of {Sharing} {Driving} {Attitude} {Information}: {A} {Quantitative} {Study} on {Lane} {Changing}}, shorttitle = {Impact of {Sharing} {Driving} {Attitude} {Information}}, doi = {10.1109/IV47402.2020.9304804}, booktitle = {2020 {IEEE} {Intelligent} {Vehicles} {Symposium} ({IV})}, author = {Liu, Xiangguo and Masoud, Neda and Zhu, Qi}, month = oct, year = {2020}, note = {ISSN: 2642-7214}, keywords = {Robots, Safety, Vehicles, Mathematical model, Trajectory, Measurement, Acceleration}, pages = {1998--2005} }
One for Many: Transfer Learning for Building HVAC Control

Shichao Xu, Yixuan Wang, Yanzhi Wang, and 2 more authors

In Proceedings of the 7th ACM International Conference on Systems for Energy-Efficient Buildings, Cities, and Transportation, Nov 2020

Abs DOI Bib HTML

The design of building heating, ventilation, and air conditioning (HVAC) system is critically important, as it accounts for around half of building energy consumption and directly affects occupant comfort, productivity, and health. Traditional HVAC control methods are typically based on creating explicit physical models for building thermal dynamics, which often require significant effort to develop and are difficult to achieve sufficient accuracy and efficiency for runtime building control and scalability for field implementations. Recently, deep reinforcement learning (DRL) has emerged as a promising data-driven method that provides good control performance without analyzing physical models at runtime. However, a major challenge to DRL (and many other data-driven learning methods) is the long training time it takes to reach the desired performance. In this work, we present a novel transfer learning based approach to overcome this challenge. Our approach can effectively transfer a DRL-based HVAC controller trained for the source building to a controller for the target building with minimal effort and improved performance, by decomposing the design of neural network controller into a transferable front-end network that captures building-agnostic behavior and a back-end network that can be efficiently trained for each specific building. We conducted experiments on a variety of transfer scenarios between buildings with different sizes, numbers of thermal zones, materials and layouts, air conditioner types, and ambient weather conditions. The experimental results demonstrated the effectiveness of our approach in significantly reducing the training time, energy cost, and temperature violations.
@inproceedings{xuOneManyTransfer2020, address = {New York, NY, USA}, series = {{BuildSys} '20}, title = {One for {Many}: {Transfer} {Learning} for {Building} {HVAC} {Control}}, isbn = {978-1-4503-8061-4}, shorttitle = {One for {Many}}, url = {https://doi.org/10.1145/3408308.3427617}, doi = {10.1145/3408308.3427617}, urldate = {2022-05-05}, booktitle = {Proceedings of the 7th {ACM} {International} {Conference} on {Systems} for {Energy}-{Efficient} {Buildings}, {Cities}, and {Transportation}}, publisher = {Association for Computing Machinery}, author = {Xu, Shichao and Wang, Yixuan and Wang, Yanzhi and O'Neill, Zheng and Zhu, Qi}, month = nov, year = {2020}, keywords = {Deep reinforcement learning, HVAC control, Data-driven, Smart Buildings, Transfer learning}, pages = {230--239} }
RTSS
GoodSpread: Criticality-Aware Static Scheduling of CPS with Multi-QoS Resources

Debayan Roy, Sumana Ghosh, Qi Zhu, and 2 more authors

In 2020 IEEE Real-Time Systems Symposium (RTSS), Dec 2020

ISSN: 2576-3172

Abs DOI Bib

In practice, safety-critical cyber-physical systems (CPS) are often implemented using high quality-of-service (QoS) resources to provide maximum performance in all scenarios. Such implementations are oblivious to the changing criticality levels of CPS based on their physical dynamics (e.g., steady or transient state). Considering that high-QoS resources are constrained for cost-sensitive CPS, such criticality-oblivious implementations are highly inefficient. Towards a tighter dimensioning of these resources, state-of-the-art approaches have considered multi-QoS resources and studied criticality-aware dynamic resource allocation along the lines of mixed-criticality systems. However, these approaches have high implementation overheads. Moreover, in safety-critical domains like automotive and avionics, certification of such dynamic policies is challenging and the implementation platforms typically do not support dynamic reconfiguration. To address these challenges, we present GoodSpread that uses a static scheduling strategy and offers the same performance guarantees while saving resources (more than 50 % in certain cases) compared to the existing dynamic schemes. The main idea here is to spread the high-QoS resources as uniformly as possible over time in order to accommodate the uncertainty of when the criticality level might change. Our proposed strategy studies the physical dynamics to determine the spread factor, i.e., how often the high-QoS resources need to be provisioned. We further propose an extensibility-driven optimization approach to obtain a static schedule that will accommodate future workloads on the remaining resources with maximum flexibility.
@inproceedings{royGoodSpreadCriticalityAwareStatic2020, title = {{GoodSpread}: {Criticality}-{Aware} {Static} {Scheduling} of {CPS} with {Multi}-{QoS} {Resources}}, shorttitle = {{GoodSpread}}, doi = {10.1109/RTSS49844.2020.00026}, booktitle = {2020 {IEEE} {Real}-{Time} {Systems} {Symposium} ({RTSS})}, author = {Roy, Debayan and Ghosh, Sumana and Zhu, Qi and Caccamo, Marco and Chakraborty, Samarjit}, month = dec, year = {2020}, note = {ISSN: 2576-3172}, keywords = {Resource management, Vehicle dynamics, Optimization, Uncertainty, Schedules, Dynamic scheduling, cyber physical systems, hybrid optimization, mixed criticality systems, multi QoS resources, Quality of service, resource efficiency, safety critical systems}, pages = {178--190} }
Addressing Class Imbalance in Federated Learning

Lixu Wang, Shichao Xu, Xiao Wang, and 1 more author

arXiv:2008.06217 [cs, stat], Dec 2020

arXiv: 2008.06217

Abs Bib HTML

Federated learning (FL) is a promising approach for training decentralized data located on local client devices while improving efficiency and privacy. However, the distribution and quantity of the training data on the clients’ side may lead to significant challenges such as class imbalance and non-IID (non-independent and identically distributed) data, which could greatly impact the performance of the common model. While much effort has been devoted to helping FL models converge when encountering non-IID data, the imbalance issue has not been sufficiently addressed. In particular, as FL training is executed by exchanging gradients in an encrypted form, the training data is not completely observable to either clients or servers, and previous methods for class imbalance do not perform well for FL. Therefore, it is crucial to design new methods for detecting class imbalance in FL and mitigating its impact. In this work, we propose a monitoring scheme that can infer the composition of training data for each FL round, and design a new loss function – }textbf{Ratio Loss} to mitigate the impact of the imbalance. Our experiments demonstrate the importance of acknowledging class imbalance and taking measures as early as possible in FL training, and the effectiveness of our method in mitigating the impact. Our method is shown to significantly outperform previous methods, while maintaining client privacy.
@article{wangAddressingClassImbalance2020, title = {Addressing {Class} {Imbalance} in {Federated} {Learning}}, url = {http://arxiv.org/abs/2008.06217}, urldate = {2022-05-06}, journal = {arXiv:2008.06217 [cs, stat]}, author = {Wang, Lixu and Xu, Shichao and Wang, Xiao and Zhu, Qi}, month = dec, year = {2020}, note = {arXiv: 2008.06217}, keywords = {Computer Science - Machine Learning, Statistics - Machine Learning} }
DAC
Opportunistic intermittent control with safety guarantees for autonomous systems

Chao Huang, Shichao Xu, Zhilu Wang, and 3 more authors

In Proceedings of the 57th ACM/EDAC/IEEE Design Automation Conference, Jul 2020

Abs Bib

Control schemes for autonomous systems are often designed in a way that anticipates the worst case in any situation. At runtime, however, there could exist opportunities to leverage the characteristics of specific environment and operation context for more efficient control. In this work, we develop an online intermittent-control framework that combines formal verification with model-based optimization and deep reinforcement learning to opportunistically skip certain control computation and actuation to save actuation energy and computational resources without compromising system safety. Experiments on an adaptive cruise control system demonstrate that our approach can achieve significant energy and computation savings.
@inproceedings{huangOpportunisticIntermittentControl2020, address = {Virtual Event, USA}, series = {{DAC} '20}, title = {Opportunistic intermittent control with safety guarantees for autonomous systems}, isbn = {978-1-4503-6725-7}, urldate = {2022-05-05}, booktitle = {Proceedings of the 57th {ACM}/{EDAC}/{IEEE} {Design} {Automation} {Conference}}, publisher = {IEEE Press}, author = {Huang, Chao and Xu, Shichao and Wang, Zhilu and Lan, Shuyue and Li, Wenchao and Zhu, Qi}, month = jul, year = {2020}, keywords = {energy saving, formal methods, opportunistic intermittent control, robust control invariant, safe RL, safety guarantee}, pages = {1--6} }
SAW: A Tool for Safety Analysis of Weakly-Hard Systems

Chao Huang, Kai-Chieh Chang, Chung-Wei Lin, and 1 more author

In Computer Aided Verification, 2020

Abs DOI Bib

We introduce SAW, a tool for safety analysis of weakly-hard systems, in which traditional hard timing constraints are relaxed to allow bounded deadline misses for improving design flexibility and runtime resiliency. Safety verification is a key issue for weakly-hard systems, as it ensures system safety under allowed deadline misses. Previous works are either for linear systems only, or limited to a certain type of nonlinear systems (e.g., systems that satisfy exponential stability and Lipschitz continuity of the system dynamics). In this work, we propose a new technique for infinite-time safety verification of general nonlinear weakly-hard systems. Our approach first discretizes the safe state set into grids and constructs a directed graph, where nodes represent the grids and edges represent the reachability relation. Based on graph theory and dynamic programming, our approach can effectively find the safe initial set (consisting of a set of grids), from which the system can be proven safe under given weakly-hard constraints. Experimental results demonstrate the effectiveness of our approach, when compared with the state-of-the-art. An open source implementation of our tool is available at https://github.com/551100kk/SAW. The virtual machine where the tool is ready to run can be found at https://www.csie.ntu.edu.tw/~r08922054/SAW.ova.
@inproceedings{huangSAWToolSafety2020, address = {Cham}, series = {Lecture {Notes} in {Computer} {Science}}, title = {{SAW}: {A} {Tool} for {Safety} {Analysis} of {Weakly}-{Hard} {Systems}}, isbn = {978-3-030-53288-8}, shorttitle = {{SAW}}, doi = {10.1007/978-3-030-53288-8_26}, language = {en}, booktitle = {Computer {Aided} {Verification}}, publisher = {Springer International Publishing}, author = {Huang, Chao and Chang, Kai-Chieh and Lin, Chung-Wei and Zhu, Qi}, editor = {Lahiri, Shuvendu K. and Wang, Chao}, year = {2020}, keywords = {Graph theory, Safety verification, Weakly-hard systems}, pages = {543--555} }
ReachNN*: A Tool for Reachability Analysis of Neural-Network Controlled Systems

Jiameng Fan, Chao Huang, Xin Chen, and 2 more authors

In Automated Technology for Verification and Analysis, 2020

Abs DOI Bib

We introduce ReachNN*, a tool for reachability analysis of neural-network controlled systems (NNCSs). The theoretical foundation of ReachNN* is the use of Bernstein polynomials to approximate any Lipschitz-continuous neural-network controller with different types of activation functions, with provable approximation error bounds. In addition, the sampling-based error bound estimation in ReachNN* is amenable to GPU-based parallel computing. For further improvement in runtime and error bound estimation, ReachNN* also features optional controller re-synthesis via a technique called verification-aware knowledge distillation (KD) to reduce the Lipschitz constant of the neural-network controller. Experiment results across a set of benchmarks show \\7}times \\to \\422}times \\efficiency improvement over the previous prototype. Moreover, KD enables proof of reachability of NNCSs whose verification results were previously unknown due to large overapproximation errors. An open-source implementation of ReachNN* is available at https://github.com/JmfanBU/ReachNNStar.git.
@inproceedings{fanReachNNToolReachability2020, address = {Cham}, series = {Lecture {Notes} in {Computer} {Science}}, title = {{ReachNN}*: {A} {Tool} for {Reachability} {Analysis} of {Neural}-{Network} {Controlled} {Systems}}, isbn = {978-3-030-59152-6}, shorttitle = {{ReachNN}*}, doi = {10.1007/978-3-030-59152-6_30}, language = {en}, booktitle = {Automated {Technology} for {Verification} and {Analysis}}, publisher = {Springer International Publishing}, author = {Fan, Jiameng and Huang, Chao and Chen, Xin and Li, Wenchao and Zhu, Qi}, editor = {Hung, Dang Van and Sokolsky, Oleg}, year = {2020}, keywords = {Bernstein polynomials, GPU acceleration, Knowledge distillation., Neural-network controlled systems, Reachability}, pages = {537--542} }
Divide and Slide: Layer-Wise Refinement for Output Range Analysis of Deep Neural Networks

Chao Huang, Jiameng Fan, Xin Chen, and 2 more authors

IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems, Nov 2020

Conference Name: IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems

Abs DOI Bib

In this article, we present a layer-wise refinement method for neural network output range analysis. While approaches such as nonlinear programming (NLP) can directly model the high nonlinearity brought by neural networks in output range analysis, they are known to be difficult to solve in general. We propose to use a convex polygonal relaxation (overapproximation) of the activation functions to cope with the nonlinearity. This allows us to encode the relaxed problem into a mixedinteger linear program (MILP), and control the tightness of the relaxation by adjusting the number of segments in the polygon. Starting with a segment number of 1 for each neuron, which coincides with a linear programming (LP) relaxation, our approach selects neurons layer by layer to iteratively refine this relaxation. To tackle the increase of the number of integer variables with tighter refinement, we bridge the propagation-based method and the programming-based method by dividing and sliding the layerwise constraints. Specifically, given a sliding number s, for the neurons in layer l, we only encode the constraints of the layers between l - s and l. We show that our overall framework is sound and provides a valid overapproximation. Experiments on deep neural networks demonstrate significant improvement on output range analysis precision using our approach compared to the state-of-the-art.
@article{huangDivideSlideLayerWise2020, title = {Divide and {Slide}: {Layer}-{Wise} {Refinement} for {Output} {Range} {Analysis} of {Deep} {Neural} {Networks}}, volume = {39}, issn = {1937-4151}, shorttitle = {Divide and {Slide}}, doi = {10.1109/TCAD.2020.3013071}, number = {11}, journal = {IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems}, author = {Huang, Chao and Fan, Jiameng and Chen, Xin and Li, Wenchao and Zhu, Qi}, month = nov, year = {2020}, note = {Conference Name: IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems}, keywords = {neural networks, Estimation, Neurons, Programming, Linear programming, Biological neural networks, Convolution, Linear programming (LP), Microsoft Windows, mixed-integer linear programming (MILP), output range analysis, refinement}, pages = {3323--3335} }
Distributed Multi-agent Video Fast-forwarding

Shuyue Lan, Zhilu Wang, Amit K. Roy-Chowdhury, and 2 more authors

In Proceedings of the 28th ACM International Conference on Multimedia, Oct 2020

Abs Bib HTML

In many intelligent systems, a network of agents collaboratively perceives the environment for better and more efficient situation awareness. As these agents often have limited resources, it could be greatly beneficial to identify the content overlapping among camera views from different agents and leverage it for reducing the processing, transmission and storage of redundant/unimportant video frames. This paper presents a consensus-based distributed multi-agent video fast-forwarding framework, named DMVF, that fast-forwards multi-view video streams collaboratively and adaptively. In our framework, each camera view is addressed by a reinforcement learning based fast-forwarding agent, which periodically chooses from multiple strategies to selectively process video frames and transmits the selected frames at adjustable paces. During every adaptation period, each agent communicates with a number of neighboring agents, evaluates the importance of the selected frames from itself and those from its neighbors, refines such evaluation together with other agents via a system-wide consensus algorithm, and uses such evaluation to decide their strategy for the next period. Compared with approaches in the literature on a real-world surveillance video dataset VideoWeb, our method significantly improves the coverage of important frames and also reduces the number of frames processed in the system.
@incollection{lanDistributedMultiagentVideo2020, address = {New York, NY, USA}, title = {Distributed {Multi}-agent {Video} {Fast}-forwarding}, isbn = {978-1-4503-7988-5}, url = {https://doi.org/10.1145/3394171.3413767}, urldate = {2022-05-05}, booktitle = {Proceedings of the 28th {ACM} {International} {Conference} on {Multimedia}}, publisher = {Association for Computing Machinery}, author = {Lan, Shuyue and Wang, Zhilu and Roy-Chowdhury, Amit K. and Wei, Ermin and Zhu, Qi}, month = oct, year = {2020}, keywords = {distributed optimization, multi-agent, video fast-forwarding}, pages = {1075--1084} }
ICCAD
Leveraging Weakly-hard Constraints for Improving System Fault Tolerance with Functional and Timing Guarantees

Hengyi Liang, Zhilu Wang, Ruochen Jiao, and 1 more author

In 2020 IEEE/ACM International Conference On Computer Aided Design (ICCAD), Nov 2020

ISSN: 1558-2434

Abs Bib

Many safety-critical real-time systems operate under harsh environment and are subject to soft errors caused by transient or intermittent faults. It is critical and yet often very challenging to apply fault tolerance techniques in these systems, due to resource limitations and stringent constraints on timing and functionality. In this work, we leverage the concept of weakly-hard constraints, which allows task deadline misses in a bounded manner, to improve system’s capability to accommodate fault tolerance techniques while ensuring timing and functional correctness. In particular, we a) quantitatively measure control cost under different deadline hit/miss scenarios and identify weak-hard constraints that guarantee control stability; b) employ typical worst-case analysis (TWCA) to bound the number of deadline misses and approximate system control cost; c) develop an event-based simulation method to check the task execution pattern and evaluate system control cost for any given solution; and d) develop a meta-heuristic algorithm that consists of heuristic methods and a simulated annealing procedure to explore the design space. Our experiments on an industrial case study and synthetic examples demonstrate the effectiveness of our approach.
@inproceedings{liangLeveragingWeaklyhardConstraints2020, title = {Leveraging {Weakly}-hard {Constraints} for {Improving} {System} {Fault} {Tolerance} with {Functional} and {Timing} {Guarantees}}, booktitle = {2020 {IEEE}/{ACM} {International} {Conference} {On} {Computer} {Aided} {Design} ({ICCAD})}, author = {Liang, Hengyi and Wang, Zhilu and Jiao, Ruochen and Zhu, Qi}, month = nov, year = {2020}, note = {ISSN: 1558-2434}, keywords = {Task analysis, Stability analysis, Fault tolerance, Fault tolerant systems, Timing, weakly-hard, Real-time systems, Analytical models, EED, EOC, timing guarantees, Transient analysis}, pages = {1--9} }
Know the unknowns: addressing disturbances and uncertainties in autonomous systems

Qi Zhu, Wenchao Li, Hyoseung Kim, and 8 more authors

In Proceedings of the 39th International Conference on Computer-Aided Design, Dec 2020

Abs DOI Bib HTML

Future autonomous systems will employ complex sensing, computation, and communication components for their perception, planning, control, and coordination, and could operate in highly dynamic and uncertain environment with safety and security assurance. To realize this vision, we have to better understand and address the challenges from the "unknowns" - the unexpected disturbances from component faults, environmental interference, and malicious attacks, as well as the inherent uncertainties in system inputs, model inaccuracies, and machine learning techniques (particularly those based on neural networks). In this work, we will discuss these challenges, propose our approaches in addressing them, and present some of the initial results. In particular, we will introduce a cross-layer framework for modeling and mitigating execution uncertainties (e.g., timing violations, soft errors) with weakly-hard paradigm, quantitative and formal methods for ensuring safe and time-predictable application of neural networks in both perception and decision making, and safety-assured adaptation strategies in dynamic environment.
@inproceedings{zhuKnowUnknownsAddressing2020a, address = {New York, NY, USA}, series = {{ICCAD} '20}, title = {Know the unknowns: addressing disturbances and uncertainties in autonomous systems}, isbn = {978-1-4503-8026-3}, shorttitle = {Know the unknowns}, url = {https://dl.acm.org/doi/10.1145/3400302.3415768}, doi = {10.1145/3400302.3415768}, urldate = {2023-04-04}, booktitle = {Proceedings of the 39th {International} {Conference} on {Computer}-{Aided} {Design}}, publisher = {Association for Computing Machinery}, author = {Zhu, Qi and Li, Wenchao and Kim, Hyoseung and Xiang, Yecheng and Wardega, Kacper and Wang, Zhilu and Wang, Yixuan and Liang, Hengyi and Huang, Chao and Fan, Jiameng and Choi, Hyunjong}, month = dec, year = {2020}, keywords = {Task analysis, adaptation, neural networks, Safety, Wireless communication, Uncertainty, Autonomous systems, disturbance, safety verification, Software, Timing, uncertainty, weakly-hard, autonomous systems}, pages = {1--9} }

2019

Application level attacks on connected vehicle protocols: 22nd International Symposium on Research in Attacks, Intrusions and Defenses, RAID 2019

Ahmed Abdo, Sakib Md Bin Malek, Zhiyun Qian, and 3 more authors

RAID 2019 Proceedings - 22nd International Symposium on Research in Attacks, Intrusions and Defenses, 2019

Publisher: USENIX Association

Abs Bib HTML

Connected vehicles (CV) applications are an emerging new technology that promises to revolutionize transportation systems. CV applications can improve safety, efficiency, and capacity of transportation systems while reducing their environmental footprints. A large number of CV applications have been proposed towards these goals, with the US Department of Transportation (US DOT) recently initiating three deployment sites. Unfortunately, the security of these protocols has not been considered carefully, and due to the fact that they affect the control of vehicles, vulnerabilities can lead to breakdowns in safety (causing accidents), performance (causing congestion and reducing capacity), or fairness (vehicles cheating the intersection management system). In this paper, we perform a detailed analysis of a recently published CV-based application protocol, Cooperative Adaptive Cruise Control (CACC), and use this analysis to classify the types of vulnerabilities that occur in the context of connected Cyber-physical systems such as CV. We show using simulations that these attacks can be extremely dangerous: we illustrate attacks that cause crashes or stall emergency vehicles. We also carry out a more systematic analysis of the impact of the attacks showing that even an individual attacker can have substantial effects on traffic flow and safety even in the presence of message security standard developed by US DOT. We believe that these attacks can be carried over to other CV applications if they are not carefully designed. The paper also explores a defense framework to mitigate these classes of vulnerabilities in CV applications.
@article{abdoApplicationLevelAttacks2019, series = {{RAID} 2019 {Proceedings} - 22nd {International} {Symposium} on {Research} in {Attacks}, {Intrusions} and {Defenses}}, title = {Application level attacks on connected vehicle protocols: 22nd {International} {Symposium} on {Research} in {Attacks}, {Intrusions} and {Defenses}, {RAID} 2019}, shorttitle = {Application level attacks on connected vehicle protocols}, url = {http://www.scopus.com/inward/record.url?scp=85103438630&partnerID=8YFLogxK}, urldate = {2022-05-06}, journal = {RAID 2019 Proceedings - 22nd International Symposium on Research in Attacks, Intrusions and Defenses}, author = {Abdo, Ahmed and Malek, Sakib Md Bin and Qian, Zhiyun and Zhu, Qi and Barth, Matthew and Abu-Ghazaleh, Nael}, year = {2019}, note = {Publisher: USENIX Association}, pages = {459--471} }
ReachNN: Reachability Analysis of Neural-Network Controlled Systems

Chao Huang, Jiameng Fan, Wenchao Li, and 2 more authors

ACM Transactions on Embedded Computing Systems, Oct 2019

Abs DOI Bib HTML

Applying neural networks as controllers in dynamical systems has shown great promises. However, it is critical yet challenging to verify the safety of such control systems with neural-network controllers in the loop. Previous methods for verifying neural network controlled systems are limited to a few specific activation functions. In this work, we propose a new reachability analysis approach based on Bernstein polynomials that can verify neural-network controlled systems with a more general form of activation functions, i.e., as long as they ensure that the neural networks are Lipschitz continuous. Specifically, we consider abstracting feedforward neural networks with Bernstein polynomials for a small subset of inputs. To quantify the error introduced by abstraction, we provide both theoretical error bound estimation based on the theory of Bernstein polynomials and more practical sampling based error bound estimation, following a tight Lipschitz constant estimation approach based on forward reachability analysis. Compared with previous methods, our approach addresses a much broader set of neural networks, including heterogeneous neural networks that contain multiple types of activation functions. Experiment results on a variety of benchmarks show the effectiveness of our approach.
@article{huangReachNNReachabilityAnalysis2019, title = {{ReachNN}: {Reachability} {Analysis} of {Neural}-{Network} {Controlled} {Systems}}, volume = {18}, issn = {1539-9087}, shorttitle = {{ReachNN}}, url = {https://doi.org/10.1145/3358228}, doi = {10.1145/3358228}, number = {5s}, urldate = {2022-05-06}, journal = {ACM Transactions on Embedded Computing Systems}, author = {Huang, Chao and Fan, Jiameng and Li, Wenchao and Chen, Xin and Zhu, Qi}, month = oct, year = {2019}, keywords = {verification, Bernstein polynomials, Neural network controlled systems, reachability}, pages = {106:1--106:22} }
Model-Based Software Synthesis for Safety-Critical Cyber-Physical Systems

Bowen Zheng, Hengyi Liang, Zhilu Wang, and 1 more author

In Safe, Autonomous and Intelligent Vehicles, 2019

Abs DOI Bib HTML

In many cyber-physical systems (CPS), software has become critical and drives future innovations. CPS software development, however, faces significant challenges from increasing functional and architectural complexity, dynamic and uncertain physical environment, and diverse design objectives and stringent system requirements. In this chapter, we introduce a model-based software synthesis flow that optimizes the generation of software tasks from functional models and the mapping of those tasks onto embedded platforms, with respect to system timing, security, fault tolerance, performance, modularity, reusability, memory usage, etc. Our approach addresses timing holistically throughout task generation and task mapping, ensures functional correctness, and enables quantitative trade-offs among different design objectives.
@incollection{zhengModelBasedSoftwareSynthesis2019, address = {Cham}, series = {Unmanned {System} {Technologies}}, title = {Model-{Based} {Software} {Synthesis} for {Safety}-{Critical} {Cyber}-{Physical} {Systems}}, isbn = {978-3-319-97301-2}, url = {https://doi.org/10.1007/978-3-319-97301-2_9}, language = {en}, urldate = {2022-05-06}, booktitle = {Safe, {Autonomous} and {Intelligent} {Vehicles}}, publisher = {Springer International Publishing}, author = {Zheng, Bowen and Liang, Hengyi and Wang, Zhilu and Zhu, Qi}, editor = {Yu, Huafeng and Li, Xin and Murray, Richard M. and Ramesh, S. and Tomlin, Claire J.}, year = {2019}, doi = {10.1007/978-3-319-97301-2_9}, keywords = {Architecture Analysis And Design Language (AADL), Cyber-physical Systems (CPS), Software Synthesis, Synthesis Flow, Time-Triggered Ethernet}, pages = {163--186} }
Design and Analysis of Delay-Tolerant Intelligent Intersection Management

Bowen Zheng, Chung-Wei Lin, Shinichi Shiraishi, and 1 more author

ACM Transactions on Cyber-Physical Systems, Nov 2019

Abs DOI Bib HTML

The rapid development of vehicular network and autonomous driving technologies provides opportunities to significantly improve transportation safety and efficiency. One promising application is centralized intelligent intersection management, where an intersection manager accepts requests from approaching vehicles (via vehicle-to-infrastructure communication messages) and schedules the order for those vehicles to safely crossing the intersection. However, communication delays and packet losses may occur due to the unreliable nature of wireless communication or malicious security attacks (e.g., jamming and flooding), and could cause deadlocks and unsafe situations. In our previous work, we considered these issues and proposed a delay-tolerant intersection management protocol for intersections with a single lane in each direction. In this work, we address key challenges in efficiency and deadlock when there are multiple lanes from each direction, and propose a delay-tolerant protocol for general multi-lane intersection management. We prove that this protocol is deadlock free, safe, and satisfies the liveness property. Furthermore, we extend the traffic simulation suite SUMO with communication modules, implement our protocol in the extended simulator, and quantitatively analyze its performance with the consideration of communication delays. Finally, we also model systems that use smart traffic lights with various back-pressure scheduling methods in SUMO, including the basic back-pressure control, the capacity-aware back-pressure control, and the adaptive max-pressure control. We then compare our delay-tolerant intelligent intersection protocol with smart traffic lights that use the three back-pressure scheduling methods, in the case of a network of interconnected intersections. Simulation results demonstrate that our approach significant outperforms the smart traffic lights under normal operation (i.e., when the communication delay is not too large).
@article{zhengDesignAnalysisDelayTolerant2019, title = {Design and {Analysis} of {Delay}-{Tolerant} {Intelligent} {Intersection} {Management}}, volume = {4}, issn = {2378-962X}, url = {https://doi.org/10.1145/3300184}, doi = {10.1145/3300184}, number = {1}, urldate = {2022-05-06}, journal = {ACM Transactions on Cyber-Physical Systems}, author = {Zheng, Bowen and Lin, Chung-Wei and Shiraishi, Shinichi and Zhu, Qi}, month = nov, year = {2019}, keywords = {vehicular network, V2X, autonomous intersection, back-pressure control, connected and autonomous vehicles, Cyber-physical security, smart traffic lights, timing attack}, pages = {3:1--3:27} }
RTAS
Job-Class-Level Fixed Priority Scheduling of Weakly-Hard Real-Time Systems

Hyunjong Choi, Hyoseung Kim, and Qi Zhu

In 2019 IEEE Real-Time and Embedded Technology and Applications Symposium (RTAS), Apr 2019

ISSN: 2642-7346

Abs DOI Bib

Many cyber-physical applications including sensing and control operations can tolerate a certain degree of timing violations as long as the number of the violations are predictably bounded. The notion of weakly-hard real-time systems has been studied to capture this effect, but existing work reveals limitations for practical use due the restrictions imposed on timing model and the high complexity of analysis. In this paper, we propose a new job-class-level fixed-priority preemptive scheduler and its schedulability analysis framework for sporadic tasks with weakly-hard real-time constraints. Our proposed scheduler employs the meet-oriented classification of jobs of a task in order to reduce the worst-case temporal interference imposed on other tasks. Under this approach, each job is associated with a "job-class" that is determined by the number of deadlines previously met (with a bounded number of consecutively-missed deadlines). This approach also allows decomposing the complex weakly-hard schedulability problem into two sub-problems that are easier to solve: (1) analyzing the response time of a job with each job-class, which can be done by an extension of the existing task-level analysis, and (2) finding possible job-class patterns, which can be modeled as a simple reachability tree. Experimental results indicate that our scheduler outperforms prior work in terms of task schedulability and analysis time complexity. We have also implemented a prototype of a job-class-level scheduler in the Linux kernel running on Raspberry Pi with acceptably-small runtime overhead.
@inproceedings{choiJobClassLevelFixedPriority2019, title = {Job-{Class}-{Level} {Fixed} {Priority} {Scheduling} of {Weakly}-{Hard} {Real}-{Time} {Systems}}, doi = {10.1109/RTAS.2019.00028}, booktitle = {2019 {IEEE} {Real}-{Time} and {Embedded} {Technology} and {Applications} {Symposium} ({RTAS})}, author = {Choi, Hyunjong and Kim, Hyoseung and Zhu, Qi}, month = apr, year = {2019}, note = {ISSN: 2642-7346}, keywords = {Task analysis, Complexity theory, Timing, Real-time systems, cyber-physical systems, Schedules, Dynamic scheduling, Jitter, real-time scheduling, weakly-hard real-time system}, pages = {241--253} }
Formal verification of weakly-hard systems

Chao Huang, Wenchao Li, and Qi Zhu

In Proceedings of the 22nd ACM International Conference on Hybrid Systems: Computation and Control, Apr 2019

Abs DOI Bib HTML

Weakly-hard systems are real-time systems that can tolerate occasional deadline misses in a bounded manner. Compared with traditional systems with hard deadline constraints, they provide more scheduling flexibility, and thus expand the design space for system configuration and reconfiguration. A key question for such a system is precisely to what degree it can tolerate deadline misses while still meeting its functional requirements. In this paper, we provide a formal treatment to the verification problem of a general class of weakly-hard systems. We discuss relaxation and over-approximation techniques for managing the complexity of reachability analysis, and develop algorithms based upon these for verifying the safety of weakly-hard systems. Experiments demonstrate the effectiveness of our approach in understanding the impact of and guiding the selection among different weakly-hard constraints.
@inproceedings{huangFormalVerificationWeaklyhard2019, address = {New York, NY, USA}, series = {{HSCC} '19}, title = {Formal verification of weakly-hard systems}, isbn = {978-1-4503-6282-5}, url = {https://doi.org/10.1145/3302504.3311811}, doi = {10.1145/3302504.3311811}, urldate = {2022-05-05}, booktitle = {Proceedings of the 22nd {ACM} {International} {Conference} on {Hybrid} {Systems}: {Computation} and {Control}}, publisher = {Association for Computing Machinery}, author = {Huang, Chao and Li, Wenchao and Zhu, Qi}, month = apr, year = {2019}, keywords = {safety, formal verification, weakly-hard}, pages = {197--207} }
ICCD
Security-Driven Codesign with Weakly-Hard Constraints for Real-Time Embedded Systems

Hengyi Liang, Zhilu Wang, Debayan Roy, and 3 more authors

In 2019 IEEE 37th International Conference on Computer Design (ICCD), Nov 2019

ISSN: 2576-6996

Abs DOI Bib

For many embedded systems, such as automotive electronic systems, security has become a pressing challenge. Limited resources and tight timing constraints often make it difficult to apply even lightweight authentication and intrusion detection schemes, especially when retrofitting existing designs. Moreover, traditional hard deadline assumption is insufficient to describe control tasks that have certain degrees of robustness and can tolerate some deadline misses while satisfying functional properties such as stability. In this work, we explore feasible weakly-hard constraints on control tasks, and then leverage the scheduling flexibility from those allowed misses to enhance system’s capability for accommodating security monitoring tasks. We develop a co-design approach that 1) sets feasible weakly-hard constraints on control tasks based on quantitative analysis, ensuring the satisfaction of control stability and performance requirements; and 2) optimizes the allocation, priority, and period assignment of security monitoring tasks, improving system security while meeting timing constraints (including the weakly-hard constraints on control tasks). Experimental results on an industrial case study and a set of synthetic examples demonstrated the significant potential of leveraging weakly-hard constraints to improve security and the effectiveness of our approach in exploring the design space to fully realize such potential.
@inproceedings{liangSecurityDrivenCodesignWeaklyHard2019, title = {Security-{Driven} {Codesign} with {Weakly}-{Hard} {Constraints} for {Real}-{Time} {Embedded} {Systems}}, doi = {10.1109/ICCD46524.2019.00035}, booktitle = {2019 {IEEE} 37th {International} {Conference} on {Computer} {Design} ({ICCD})}, author = {Liang, Hengyi and Wang, Zhilu and Roy, Debayan and Dey, Soumyajit and Chakraborty, Samarjit and Zhu, Qi}, month = nov, year = {2019}, note = {ISSN: 2576-6996}, keywords = {Task analysis, weakly hard, control, Intrusion detection, Stability analysis, Timing, Space exploration, Monitoring, security, real time}, pages = {217--226} }
ICCAD
Towards Verification-Aware Knowledge Distillation for Neural-Network Controlled Systems: Invited Paper

Jiameng Fan, Chao Huang, Wenchao Li, and 2 more authors

In 2019 IEEE/ACM International Conference on Computer-Aided Design (ICCAD), Nov 2019

ISSN: 1558-2434

Abs DOI Bib

Neural networks are widely used in many applications ranging from classification to control. While these networks are composed of simple arithmetic operations, they are challenging to formally verify for properties such as reachability due to the presence of nonlinear activation functions. In this paper, we make the observation that Lipschitz continuity of a neural network not only can play a major role in the construction of reachable sets for neural-network controlled systems but also can be systematically controlled during training of the neural network. We build on this observation to develop a novel verification-aware knowledge distillation framework that transfers the knowledge of a trained network to a new and easier-to-verify network. Experimental results show that our method can substantially improve reachability analysis of neural-network controlled systems for several state-of-the-art tools.
@inproceedings{fanVerificationAwareKnowledgeDistillation2019, title = {Towards {Verification}-{Aware} {Knowledge} {Distillation} for {Neural}-{Network} {Controlled} {Systems}: {Invited} {Paper}}, shorttitle = {Towards {Verification}-{Aware} {Knowledge} {Distillation} for {Neural}-{Network} {Controlled} {Systems}}, doi = {10.1109/ICCAD45719.2019.8942059}, booktitle = {2019 {IEEE}/{ACM} {International} {Conference} on {Computer}-{Aided} {Design} ({ICCAD})}, author = {Fan, Jiameng and Huang, Chao and Li, Wenchao and Chen, Xin and Zhu, Qi}, month = nov, year = {2019}, note = {ISSN: 1558-2434}, keywords = {Training, Control systems, Artificial neural networks, Biological neural networks, Knowledge engineering, Tools}, pages = {1--8} }
CPS
Exploring weakly-hard paradigm for networked systems

Chao Huang, Kacper Wardega, Wenchao Li, and 1 more author

In Proceedings of the Workshop on Design Automation for CPS and IoT, Apr 2019

Abs DOI Bib HTML

Networked systems have shown great promises in various cyber-physical applications, such as automotive and transportation systems, smart buildings and infrastructures, and robotic systems. As these systems employ advanced components and interact closely with the dynamic environment, they are often subject to significant disturbances from environment interference, security attacks, and device faults. To ensure system safety, performance and other properties, it is critical to capture these disturbances and reason about their impact at the network level. In this work, we propose to use weakly-hard constraints to specify the disturbances in a bounded manner, and leverage them to formally reason about system properties. We will first present two case studies that demonstrate the impact of disturbances on various properties in networked systems and motivate the usage of weakly-hard constraints. We will then discuss several possible research directions in applying weakly-hard constraints to networked systems.
@inproceedings{huangExploringWeaklyhardParadigm2019, address = {New York, NY, USA}, series = {{DESTION} '19}, title = {Exploring weakly-hard paradigm for networked systems}, isbn = {978-1-4503-6699-1}, url = {https://dl.acm.org/doi/10.1145/3313151.3313165}, doi = {10.1145/3313151.3313165}, urldate = {2023-04-05}, booktitle = {Proceedings of the {Workshop} on {Design} {Automation} for {CPS} and {IoT}}, publisher = {Association for Computing Machinery}, author = {Huang, Chao and Wardega, Kacper and Li, Wenchao and Zhu, Qi}, month = apr, year = {2019}, pages = {51--59} }

2018

Peak-Aware Online Economic Dispatching for Microgrids

Ying Zhang, Mohammad H. Hajiesmaili, Sinan Cai, and 2 more authors

IEEE Transactions on Smart Grid, Jan 2018

Conference Name: IEEE Transactions on Smart Grid

Abs DOI Bib

By employing local renewable energy sources and power generation units while connected to the central grid, microgrid can usher in great benefits in terms of cost efficiency, power reliability, and environmental awareness. Economic dispatching is a central problem in microgrid operation, which aims at effectively scheduling various energy sources to minimize the operating cost while satisfying the electricity demand. Designing intelligent economic dispatching strategies for microgrids; however, it is drastically different from that for conventional central grids due to two unique challenges. First, the demand and renewable generation uncertainty emphasizes the need for online algorithms. Second, the widely-adopted peak-based pricing scheme brings out the need for new peak-aware strategy design. In this paper, we tackle these critical challenges and devise peak-aware online economic dispatching algorithms. We prove that our deterministic and randomized algorithms achieve the best possible competitive ratios 2 - β and e/(e - 1 + β) in the fast responding generator scenario, where β ∈ [0, 1] is the ratio between the minimum grid spot price and the local-generation price. By extensive empirical evaluations using real-world traces, we show that our online algorithms achieve near offline-optimal performance. In a representative scenario, our algorithm achieves 17.5% and 9.24% cost reduction as compared with the case without local generation units and the case using peak-oblivious algorithms, respectively.
@article{zhangPeakAwareOnlineEconomic2018, title = {Peak-{Aware} {Online} {Economic} {Dispatching} for {Microgrids}}, volume = {9}, issn = {1949-3061}, doi = {10.1109/TSG.2016.2551282}, number = {1}, journal = {IEEE Transactions on Smart Grid}, author = {Zhang, Ying and Hajiesmaili, Mohammad H. and Cai, Sinan and Chen, Minghua and Zhu, Qi}, month = jan, year = {2018}, note = {Conference Name: IEEE Transactions on Smart Grid}, keywords = {Algorithm design and analysis, Uncertainty, Economics, Microgrids, Pricing, Generators, Dispatching, economic dispatching, online algorithm, peak-aware scheduling}, pages = {323--335} }
ASP-DAC
A deep reinforcement learning framework for optimizing fuel economy of hybrid electric vehicles

Pu Zhao, Yanzhi Wang, Naehyuck Chang, and 2 more authors

In 2018 23rd Asia and South Pacific Design Automation Conference (ASP-DAC), Jan 2018

ISSN: 2153-697X

Abs DOI Bib

Hybrid electric vehicles employ a hybrid propulsion system to combine the energy efficiency of electric motor and a long driving range of internal combustion engine, thereby achieving a higher fuel economy as well as convenience compared with conventional ICE vehicles. However, the relatively complicated powertrain structures of HEVs necessitate an effective power management policy to determine the power split between ICE and EM. In this work, we propose a deep reinforcement learning framework of the HEV power management with the aim of improving fuel economy. The DRL technique is comprised of an offline deep neural network construction phase and an online deep Q-learning phase. Unlike traditional reinforcement learning, DRL presents the capability of handling the high dimensional state and action space in the actual decision-making process, making it suitable for the HEV power management problem. Enabled by the DRL technique, the derived HEV power management policy is close to optimal, fully model-free, and independent of a prior knowledge of driving cycles. Simulation results based on actual vehicle setup over real-world and testing driving cycles demonstrate the effectiveness of the proposed framework on optimizing HEV fuel economy.
@inproceedings{zhaoDeepReinforcementLearning2018, title = {A deep reinforcement learning framework for optimizing fuel economy of hybrid electric vehicles}, doi = {10.1109/ASPDAC.2018.8297305}, booktitle = {2018 23rd {Asia} and {South} {Pacific} {Design} {Automation} {Conference} ({ASP}-{DAC})}, author = {Zhao, Pu and Wang, Yanzhi and Chang, Naehyuck and Zhu, Qi and Lin, Xue}, month = jan, year = {2018}, note = {ISSN: 2153-697X}, keywords = {Batteries, Fuel economy, Hybrid electric vehicles, Ice, Power system management, Propulsion, Torque}, pages = {196--202} }
Sustainability-Oriented Evaluation and Optimization for MPSoC Task Allocation and Scheduling under Thermal and Energy Variations

Mingsong Chen, Xinqian Zhang, Haifeng Gu, and 2 more authors

IEEE Transactions on Sustainable Computing, Apr 2018

Conference Name: IEEE Transactions on Sustainable Computing

Abs DOI Bib

Aiming at high performance, more and more Cyber-Physical Systems (CPSs) adopt Multiprocessor System-on-Chips (MPSoCs) as computation units. However, due to increasing integration of transistors on a die, the power densities together with performance variations of MPSoC chips have been increasing dramatically. Consequently, the MPSoC-based CPSs might become unsustainable and unreliable. Although various Task Allocation and Scheduling (TAS) heuristics have been proposed to minimize the hotspot time (i.e., duration of thermal emergency) and energy consumption of MPSoC designs, few of them can guarantee the highest performance yield under process variations without violating energy, thermal and timing constraints. To address these challenges, this paper proposes a novel energy- and thermal-aware TAS evaluation and optimization framework. Based on statistical model checking techniques, our approach enables accurate modeling and reasoning of the performance yield of real-time MPSoC designs under joint energy and thermal constraints. To enable system-level design space exploration, we propose a regression analysis-based method that can drastically reduce the overall exploration efforts. Experimental results show that our fully-automated approach can not only allow accurate sustainability-oriented reasoning of TAS solutions under specified thermal and energy constraints, but also enable the quick search of optimal TAS solutions on different MPSoC architectures with the highest performance yield.
@article{chenSustainabilityOrientedEvaluationOptimization2018, title = {Sustainability-{Oriented} {Evaluation} and {Optimization} for {MPSoC} {Task} {Allocation} and {Scheduling} under {Thermal} and {Energy} {Variations}}, volume = {3}, issn = {2377-3782}, doi = {10.1109/TSUSC.2017.2723500}, number = {2}, journal = {IEEE Transactions on Sustainable Computing}, author = {Chen, Mingsong and Zhang, Xinqian and Gu, Haifeng and Wei, Tongquan and Zhu, Qi}, month = apr, year = {2018}, note = {Conference Name: IEEE Transactions on Sustainable Computing}, keywords = {Resource management, Reliability, Delays, Optimization, Energy consumption, optimization, Cyber-physical systems, Automata, Model checking, statistical model checking, sustainability, task allocation and scheduling}, pages = {84--97} }

CVPR

FFNet: Video Fast-Forwarding via Reinforcement Learning

Shuyue Lan, Rameswar Panda, Qi Zhu, and 1 more author

In 2018 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2018

Bib HTML

@inproceedings{lanFFNetVideoFastForwarding2018,
  booktitle = {2018 {IEEE} {Conference} on {Computer} {Vision} and {Pattern} {Recognition} ({CVPR})},
  title = {{FFNet}: {Video} {Fast}-{Forwarding} via {Reinforcement} {Learning}},
  shorttitle = {{FFNet}},
  url = {https://openaccess.thecvf.com/content_cvpr_2018/html/Lan_FFNet_Video_Fast-Forwarding_CVPR_2018_paper.html},
  urldate = {2022-05-06},
  author = {Lan, Shuyue and Panda, Rameswar and Zhu, Qi and Roy-Chowdhury, Amit K.},
  year = {2018},
  pages = {6771--6780}
}

Design Automation for Cyber-Physical Systems [Scanning the Issue]

Qi Zhu, Alberto Sangiovanni-Vincentelli, Shiyan Hu, and 1 more author

Proceedings of the IEEE, Sep 2018

Conference Name: Proceedings of the IEEE

Abs DOI Bib

Cyber-physical systems (CPSs) are characterized by the seamless integration and close interaction of cyber components (e.g., sensors, computation nodes, communication networks) and physical processes (e.g., mechanical devices, physical environment, humans). The cyber components monitor, analyze, and control the physical processes, and react to their changes through feedback loops. A classic example of CPSs is autonomous vehicles. These vehicles collect information of the surrounding physical environment via heterogeneous sensors such as cameras, radar, and LIDAR; process and analyze the multi-modal information at real time with advanced computing devices such as GPUs, application-specific SoCs and multicore CPUs; automatically make planning and control decisions; and continuously actuate the corresponding mechanical components. The cyber components of autonomous vehicles are much more intelligent and complex than those of traditional vehicles, and interact more directly and closely with the physical environment.
@article{zhuDesignAutomationCyberPhysical2018, title = {Design {Automation} for {Cyber}-{Physical} {Systems} [{Scanning} the {Issue}]}, volume = {106}, issn = {1558-2256}, doi = {10.1109/JPROC.2018.2865229}, number = {9}, journal = {Proceedings of the IEEE}, author = {Zhu, Qi and Sangiovanni-Vincentelli, Alberto and Hu, Shiyan and Li, Xin}, month = sep, year = {2018}, note = {Conference Name: Proceedings of the IEEE}, keywords = {Internet of Things, Intelligent vehicles, Autonomous automobiles, Design automation, Cyber-physical systems, Microfluidics, Network architecture, Smart buildings, Special issues and sections}, pages = {1479--1483} }
Codesign Methodologies and Tools for Cyber–Physical Systems

Qi Zhu and Alberto Sangiovanni-Vincentelli

Proceedings of the IEEE, Sep 2018

Conference Name: Proceedings of the IEEE

Abs DOI Bib

Cyber-physical system (CPS) analysis and design are challenging due to the intrinsic heterogeneity of those systems. Today, CPSs are often designed by leveraging existing solutions and by adding cyber components to an existing physical system, thus decomposing the design into two separate phases. In this paper, we argue that the codesign of the cyber and physical components would expose solutions that are better under all aspects, such as safety, efficiency, security, performance, reliability, fault tolerance, and extensibility. To do so, automated codesign tools are a necessity due to the complexity of the problems at hand. In the paper, we will discuss the key needs and challenges in developing modeling, simulation, synthesis, validation, and verification tools for CPS codesign, present promising codesign approaches from our teams and others, and point out where additional research is needed.
@article{zhuCodesignMethodologiesTools2018, title = {Codesign {Methodologies} and {Tools} for {Cyber}–{Physical} {Systems}}, volume = {106}, issn = {1558-2256}, doi = {10.1109/JPROC.2018.2864271}, number = {9}, journal = {Proceedings of the IEEE}, author = {Zhu, Qi and Sangiovanni-Vincentelli, Alberto}, month = sep, year = {2018}, note = {Conference Name: Proceedings of the IEEE}, keywords = {Mathematical model, Computational modeling, verification, Space exploration, design automation, Semantics, Codesign, Cyber-physical systems, synthesis, cyber–physical systems (CPSs), modeling, Unified modeling language}, pages = {1484--1500} }

Design Automation for Intelligent Automotive Systems

Shuyue Lan, Chao Huang, Zhilu Wang, and 3 more authors

In , Oct 2018

DOI Bib

@inproceedings{lanDesignAutomationIntelligent2018,
  title = {Design {Automation} for {Intelligent} {Automotive} {Systems}},
  doi = {10.1109/TEST.2018.8624723},
  author = {Lan, Shuyue and Huang, Chao and Wang, Zhilu and Liang, Hengyi and Su, Wenhao and Zhu, Qi},
  month = oct,
  year = {2018},
  pages = {1--10}
}

Model-based and data-driven approaches for building automation and control

Tianshu Wei, Xiaoming Chen, Xin Li, and 1 more author

In Proceedings of the International Conference on Computer-Aided Design, Nov 2018

Abs DOI Bib HTML

Smart buildings in the future are complex cyber-physical-human systems that involve close interactions among embedded platform (for sensing, computation, communication and control), mechanical components, physical environment, building architecture, and occupant activities. The design and operation of such buildings require a new set of methodologies and tools that can address these heterogeneous domains in a holistic, quantitative and automated fashion. In this paper, we will present our design automation methods for improving building energy efficiency and offering comfortable services to occupants at low cost. In particular, we will highlight our work in developing both model-based and data-driven approaches for building automation and control, including methods for co-scheduling heterogeneous energy demands and supplies, for integrating intelligent building energy management with grid optimization through a proactive demand response framework, for optimizing HVAC control with deep reinforcement learning, and for accurately measuring in-building temperature by combining prior modeling information with few sensor measurements based upon Bayesian inference.
@inproceedings{weiModelbasedDatadrivenApproaches2018, address = {New York, NY, USA}, series = {{ICCAD} '18}, title = {Model-based and data-driven approaches for building automation and control}, isbn = {978-1-4503-5950-4}, url = {https://doi.org/10.1145/3240765.3243485}, doi = {10.1145/3240765.3243485}, urldate = {2022-05-05}, booktitle = {Proceedings of the {International} {Conference} on {Computer}-{Aided} {Design}}, publisher = {Association for Computing Machinery}, author = {Wei, Tianshu and Chen, Xiaoming and Li, Xin and Zhu, Qi}, month = nov, year = {2018}, keywords = {deep reinforcement learning, Bayesian inference, data-driven, model predictive control, model-based design, smart buildings}, pages = {1--8} }
Network and system level security in connected vehicle applications

Hengyi Liang, Matthew Jagielski, Bowen Zheng, and 5 more authors

In Proceedings of the International Conference on Computer-Aided Design, Nov 2018

Abs DOI Bib HTML

Connected vehicle applications such as autonomous intersections and intelligent traffic signals have shown great promises in improving transportation safety and efficiency. However, security is a major concern in these systems, as vehicles and surrounding infrastructures communicate through ad-hoc networks. In this paper, we will first review security vulnerabilities in connected vehicle applications. We will then introduce and discuss some of the defense mechanisms at network and system levels, including (1) the Security Credential Management System (SCMS) proposed by the United States Department of Transportation, (2) an intrusion detection system (IDS) that we are developing and its application on collaborative adaptive cruise control, and (3) a partial consensus mechanism and its application on lane merging. These mechanisms can assist to improve the security of connected vehicle applications.
@inproceedings{liangNetworkSystemLevel2018, address = {New York, NY, USA}, series = {{ICCAD} '18}, title = {Network and system level security in connected vehicle applications}, isbn = {978-1-4503-5950-4}, url = {https://doi.org/10.1145/3240765.3243488}, doi = {10.1145/3240765.3243488}, urldate = {2022-05-05}, booktitle = {Proceedings of the {International} {Conference} on {Computer}-{Aided} {Design}}, publisher = {Association for Computing Machinery}, author = {Liang, Hengyi and Jagielski, Matthew and Zheng, Bowen and Lin, Chung-Wei and Kang, Eunsuk and Shiraishi, Shinichi and Nita-Rotaru, Cristina and Zhu, Qi}, month = nov, year = {2018}, keywords = {vehicular network, consensus, IDS, SCMS, security}, pages = {1--7} }

2017

SMARTCOMP
Delay-Aware Design, Analysis and Verification of Intelligent Intersection Management

Bowen Zheng, Chung-Wei Lin, Hengyi Liang, and 3 more authors

In 2017 IEEE International Conference on Smart Computing (SMARTCOMP), May 2017

Abs DOI Bib

With the rapid advancement of autonomous driving and vehicular communication technology, intelligent intersection management has shown great promise in improving transportation efficiency. In a typical intelligent intersection, an intersection manager communicates with autonomous vehicles wirelessly and schedules their crossing of the intersection. Previous system designs, however, do not address the possible communication delays due to network congestion or security attacks, and could lead to unsafe or deadlocked systems. In this work, we propose a delay- tolerant protocol for intelligent intersection management, and develop a modeling, simulation and verification framework for analyzing the protocol’s safety, liveness and performance. Experiments demonstrate the advantages of our proposed protocol over traditional traffic light control, and more importantly, demonstrate the importance and effectiveness of using this framework to address timing (delay) in vehicular network applications. This work is the first step towards a comprehensive delay-aware design and verification framework for practical vehicular network applications.
@inproceedings{zhengDelayAwareDesignAnalysis2017, title = {Delay-{Aware} {Design}, {Analysis} and {Verification} of {Intelligent} {Intersection} {Management}}, doi = {10.1109/SMARTCOMP.2017.7946999}, booktitle = {2017 {IEEE} {International} {Conference} on {Smart} {Computing} ({SMARTCOMP})}, author = {Zheng, Bowen and Lin, Chung-Wei and Liang, Hengyi and Shiraishi, Shinichi and Li, Wenchao and Zhu, Qi}, month = may, year = {2017}, keywords = {Safety, Protocols, Delays, System recovery, Analytical models, Automata, Schedules}, pages = {1--8} }
An optimal energy co-scheduling framework for smart buildings

Tiansong Cui, Shuang Chen, Yanzhi Wang, and 3 more authors

Integration, Jun 2017

Abs DOI Bib HTML

The Heating, Ventilation and Air Conditioning (HVAC) system accounts for nearly half of the energy consumption of a typical building. Additionally, the need for HVAC changes over hours and days as does the electric energy price. Level of comfort of the building occupants is, however, a primary concern, which tends to overwrite pricing. Dynamic HVAC control under a dynamic energy pricing model while meeting an acceptable level of occupants’ comfort is thus critical to achieve energy efficiency in buildings in a sustainable manner. Finally, there is the possibility that the building is equipped with some renewable sources of power such as solar panels mounted on the rooftop. The presence of Hybrid Electrical Energy Storage (HEES) system in a target building would enable peak power shaving by adopting a suitable charging and discharging schedule for each Electrical Energy Storage (EES) element, while simultaneously meeting building energy efficiency and user comfort requirements. Achieving this goal requires detailed information (or predictions) about the amount of local power generation from the renewable source plus the power consumption load of the building. This paper addresses the co-scheduling problem of HVAC control and HEES system management to achieve energy-efficient smart buildings, while also accounting for the degradation of the battery state-of-health during charging and discharging operations (which in turn determines the amortized cost of owning and utilizing a battery storage system). A time-of-use dynamic pricing scenario is assumed and various energy loss components are considered, including power dissipation in the power conversion circuitry, the rate capacity effect in the batteries, and the self-discharge in the super-capacitor. A global optimization framework targeting the entire billing cycle is presented and an adaptive co-scheduling algorithm is provided to dynamically update the optimal HVAC air flow control and the HEES system management in each time slot during the billing cycle to mitigate the prediction error of unknown parameters. Experimental results show that the proposed algorithm achieves up to 10% in the total electric utility cost compared with some baseline methods.
@article{cuiOptimalEnergyCoscheduling2017, title = {An optimal energy co-scheduling framework for smart buildings}, volume = {58}, issn = {0167-9260}, url = {https://www.sciencedirect.com/science/article/pii/S0167926016300864}, doi = {10.1016/j.vlsi.2016.10.009}, language = {en}, urldate = {2022-05-06}, journal = {Integration}, author = {Cui, Tiansong and Chen, Shuang and Wang, Yanzhi and Zhu, Qi and Nazarian, Shahin and Pedram, Massoud}, month = jun, year = {2017}, keywords = {Smart building, Heating, Ventilation and Air Conditioning (HVAC) control, Hybrid Electrical Energy Storage (HEES) system, State-of-health (SoH) Degradation}, pages = {528--537} }
Deep Reinforcement Learning for Building HVAC Control

Tianshu Wei, Yanzhi Wang, and Qi Zhu

In Proceedings of the 54th Annual Design Automation Conference 2017, Jun 2017

Abs DOI Bib HTML

Buildings account for nearly 40% of the total energy consumption in the United States, about half of which is used by the HVAC (heating, ventilation, and air conditioning) system. Intelligent scheduling of building HVAC systems has the potential to significantly reduce the energy cost. However, the traditional rule-based and model-based strategies are often inefficient in practice, due to the complexity in building thermal dynamics and heterogeneous environment disturbances. In this work, we develop a data-driven approach that leverages the deep reinforcement learning (DRL) technique, to intelligently learn the effective strategy for operating the building HVAC systems. We evaluate the performance of our DRL algorithm through simulations using the widely-adopted EnergyPlus tool. Experiments demonstrate that our DRL-based algorithm is more effective in energy cost reduction compared with the traditional rule-based approach, while maintaining the room temperature within desired range.
@inproceedings{weiDeepReinforcementLearning2017, address = {New York, NY, USA}, series = {{DAC} '17}, title = {Deep {Reinforcement} {Learning} for {Building} {HVAC} {Control}}, isbn = {978-1-4503-4927-7}, url = {https://doi.org/10.1145/3061639.3062224}, doi = {10.1145/3061639.3062224}, urldate = {2022-05-05}, booktitle = {Proceedings of the 54th {Annual} {Design} {Automation} {Conference} 2017}, publisher = {Association for Computing Machinery}, author = {Wei, Tianshu and Wang, Yanzhi and Zhu, Qi}, month = jun, year = {2017}, pages = {1--6} }
Extensibility-Driven Automotive In-Vehicle Architecture Design: Invited

Qi Zhu, Hengyi Liang, Licong Zhang, and 3 more authors

In Proceedings of the 54th Annual Design Automation Conference 2017, Jun 2017

Abs DOI Bib HTML

Increasingly more software-based applications are being developed and deployed in modern vehicles. As a result, the extensibility of a system design has become an important issue in order to accommodate more future applications and update of existing ones on one hand and reduce the effort and cost of re-design, test and validation on the other. In this paper, we discuss the extensibility-driven design in the automotive E/E architecture. We explain the motivation for such a design objective and discuss the definition of extensibility metric and extensibility-driven design methods under two different setting, namely the system based on CAN bus and FlexRay bus. Based on these two examples, we illustrate the importance and advantages of extensibility-driven design in the automotive E/E architecture.
@inproceedings{zhuExtensibilityDrivenAutomotiveInVehicle2017, address = {New York, NY, USA}, series = {{DAC} '17}, title = {Extensibility-{Driven} {Automotive} {In}-{Vehicle} {Architecture} {Design}: {Invited}}, isbn = {978-1-4503-4927-7}, shorttitle = {Extensibility-{Driven} {Automotive} {In}-{Vehicle} {Architecture} {Design}}, url = {https://doi.org/10.1145/3061639.3072956}, doi = {10.1145/3061639.3072956}, urldate = {2022-05-05}, booktitle = {Proceedings of the 54th {Annual} {Design} {Automation} {Conference} 2017}, publisher = {Association for Computing Machinery}, author = {Zhu, Qi and Liang, Hengyi and Zhang, Licong and Roy, Debayan and Li, Wenchao and Chakraborty, Samarjit}, month = jun, year = {2017}, keywords = {CAN, FlexRay, extensibility, automotive E/E architecture}, pages = {1--6} }
Design Automation of Cyber-Physical Systems: Challenges, Advances, and Opportunities

Sanjit A. Seshia, Shiyan Hu, Wenchao Li, and 1 more author

IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems, Sep 2017

Conference Name: IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems

Abs DOI Bib

A cyber-physical system (CPS) is an integration of computation with physical processes whose behavior is defined by both computational and physical parts of the system. In this paper, we present a view of the challenges and opportunities for design automation of CPS. We identify a combination of characteristics that define the challenges unique to the design automation of CPS. We then present selected promising advances in depth, focusing on four foundational directions: combining model-based and data-driven design methods; design for human-in-the-loop systems; component-based design with contracts, and design for security and privacy. These directions are illustrated with examples from two application domains: smart energy systems and next-generation automotive systems.
@article{seshiaDesignAutomationCyberPhysical2017, title = {Design {Automation} of {Cyber}-{Physical} {Systems}: {Challenges}, {Advances}, and {Opportunities}}, volume = {36}, issn = {1937-4151}, shorttitle = {Design {Automation} of {Cyber}-{Physical} {Systems}}, doi = {10.1109/TCAD.2016.2633961}, number = {9}, journal = {IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems}, author = {Seshia, Sanjit A. and Hu, Shiyan and Li, Wenchao and Zhu, Qi}, month = sep, year = {2017}, note = {Conference Name: IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems}, keywords = {privacy, Vehicle dynamics, machine learning, Computational modeling, formal verification, Design automation, Design methodology, design automation, Integrated circuit modeling, Solid modeling, security, Cyber-physical systems, automotive engineering, energy management, formal specification, human-robot interaction, synthesis}, pages = {1421--1434} }
Addressing Extensibility and Fault Tolerance in CAN-based Automotive Systems

Hengyi Liang, Zhilu Wang, Bowen Zheng, and 1 more author

In Proceedings of the Eleventh IEEE/ACM International Symposium on Networks-on-Chip, Oct 2017

Abs DOI Bib HTML

The design of automotive electronic systems needs to address a variety of important objectives, including safety, performance, fault tolerance, reliability, security, extensibility, etc. To obtain a feasible design, timing constraints must be satisfied and latencies of certain functional paths should not exceed their deadlines. From functionality perspective, soft errors caused by transient or intermittent faults need to be detected and recovered with fault tolerance techniques. Moreover, during the lifetime of a vehicle design or even the same car, updates are often needed to add new features or fix bugs in existing ones. It is therefore critical to improve the design extensibility for accommodating such updates without incurring major redesign and re-verification cost. In this work, we discuss the metrics for measuring latency, fault tolerance and extensibility, and present a simulated annealing based algorithm to search the design space with respect to them. Experimental results on industrial and synthetic examples demonstrate clear trade-offs among these objectives, and hence the importance of quantitatively analyzing such trade-offs and exploring the design space with automation tools.
@inproceedings{liangAddressingExtensibilityFault2017, address = {New York, NY, USA}, series = {{NOCS} '17}, title = {Addressing {Extensibility} and {Fault} {Tolerance} in {CAN}-based {Automotive} {Systems}}, isbn = {978-1-4503-4984-0}, url = {https://doi.org/10.1145/3130218.3130233}, doi = {10.1145/3130218.3130233}, urldate = {2022-05-05}, booktitle = {Proceedings of the {Eleventh} {IEEE}/{ACM} {International} {Symposium} on {Networks}-on-{Chip}}, publisher = {Association for Computing Machinery}, author = {Liang, Hengyi and Wang, Zhilu and Zheng, Bowen and Zhu, Qi}, month = oct, year = {2017}, pages = {1--8} }
Deep reinforcement learning: framework, applications, and embedded implementations

Hongjia Li, Tianshu Wei, Ao Ren, and 2 more authors

In Proceedings of the 36th International Conference on Computer-Aided Design, Nov 2017

Abs Bib

The recent breakthroughs of deep reinforcement learning (DRL) technique in Alpha Go and playing Atari have set a good example in handling large state and actions spaces of complicated control problems. The DRL technique is comprised of (i) an offline deep neural network (DNN) construction phase, which derives the correlation between each state-action pair of the system and its value function, and (ii) an online deep Q-learning phase, which adaptively derives the optimal action and updates value estimates. In this paper, we first present the general DRL framework, which can be widely utilized in many applications with different optimization objectives. This is followed by the introduction of three specific applications: the cloud computing resource allocation problem, the residential smart grid task scheduling problem, and building HVAC system optimal control problem. The effectiveness of the DRL technique in these three cyber-physical applications have been validated. Finally, this paper investigates the stochastic computing-based hardware implementations of the DRL framework, which consumes a significant improvement in area efficiency and power consumption compared with binary-based implementation counterparts.
@inproceedings{liDeepReinforcementLearning2017, address = {Irvine, California}, series = {{ICCAD} '17}, title = {Deep reinforcement learning: framework, applications, and embedded implementations}, shorttitle = {Deep reinforcement learning}, urldate = {2022-05-05}, booktitle = {Proceedings of the 36th {International} {Conference} on {Computer}-{Aided} {Design}}, publisher = {IEEE Press}, author = {Li, Hongjia and Wei, Tianshu and Ren, Ao and Zhu, Qi and Wang, Yanzhi}, month = nov, year = {2017}, keywords = {deep reinforcement learning, cyber-physical systems, optimal control, stochastic computing}, pages = {847--854} }
Timing and security analysis of VANET-based intelligent transportation systems

Bowen Zheng, Muhammed O. Sayin, Chung-Wei Lin, and 2 more authors

In Proceedings of the 36th International Conference on Computer-Aided Design, Nov 2017

Abs Bib

With the fast development of autonomous driving and vehicular communication technologies, intelligent transportation systems that are based on VANET (Vehicular Ad-Hoc Network) have shown great promise. For instance, through V2V (Vehicle-to-Vehicle) and V2I (Vehicle-to-Infrastructure) communication, intelligent intersections allow more fine-grained control of vehicle crossings and significantly enhance traffic efficiency. However, the performance and safety of these VANET-based systems could be seriously impaired by communication delays and packet losses, which may be caused by network congestion or by malicious attacks that target communication timing behavior. In this paper, we quantitatively model and analyze some of the timing and security issues in transportation networks with VANET-based intelligent intersections. In particular, we demonstrate how communication delays may affect the performance and safety of a single intersection and of multiple interconnected intersections, and present our delay-tolerant intersection management protocols. We also discuss the issues of such protocols when the vehicles are non-cooperative and how they may be addressed with game theory.
@inproceedings{zhengTimingSecurityAnalysis2017, address = {Irvine, California}, series = {{ICCAD} '17}, title = {Timing and security analysis of {VANET}-based intelligent transportation systems}, urldate = {2022-05-05}, booktitle = {Proceedings of the 36th {International} {Conference} on {Computer}-{Aided} {Design}}, publisher = {IEEE Press}, author = {Zheng, Bowen and Sayin, Muhammed O. and Lin, Chung-Wei and Shiraishi, Shinichi and Zhu, Qi}, month = nov, year = {2017}, keywords = {security, intelligent transportation systems, timing, VANET}, pages = {984--991} }
Quantitative Performance Evaluation of Uncertainty-Aware Hybrid AADL Designs Using Statistical Model Checking

Yongxiang Bao, Mingsong Chen, Qi Zhu, and 3 more authors

IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems, Dec 2017

Conference Name: IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems

Abs DOI Bib

The hybrid architecture analysis and design language (AADL) has been proposed to model the interactions between embedded control systems and continuous physical environment. However, the worst-case performance analysis of hybrid AADL designs often leads to overly pessimistic estimations, and is not suitable for accurate reasoning about overall system performance, in particular when the system closely interacts with an uncertain external environment. To address this challenge, this paper proposes a statistical model checking-based framework that can perform quantitative evaluation of uncertainty-aware hybrid AADL designs against various performance queries. Our approach extends hybrid AADL to support the modeling of environment uncertainties. Furthermore, we propose a set of transformation rules that can automatically translate AADL designs together with designers’ requirements into networks of priced timed automata and performance queries, respectively. Comprehensive experimental results on the movement authority scenario of Chinese train control system level 3 demonstrate the effectiveness of our approach.
@article{baoQuantitativePerformanceEvaluation2017, title = {Quantitative {Performance} {Evaluation} of {Uncertainty}-{Aware} {Hybrid} {AADL} {Designs} {Using} {Statistical} {Model} {Checking}}, volume = {36}, issn = {1937-4151}, doi = {10.1109/TCAD.2017.2681076}, number = {12}, journal = {IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems}, author = {Bao, Yongxiang and Chen, Mingsong and Zhu, Qi and Wei, Tongquan and Mallet, Frederic and Zhou, Tingliang}, month = dec, year = {2017}, note = {Conference Name: IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems}, keywords = {Computer architecture, Computational modeling, Uncertainty, uncertainty, Ports (Computers), Analytical models, Hybrid architecture analysis and design language (AADL), Model checking, quantitative performance evaluation, Statistical analysis, statistical model checking (SMC)}, pages = {1989--2002} }

2016

ASP-DAC
Optimal co-scheduling of HVAC control and battery management for energy-efficient buildings considering state-of-health degradation

Tiansong Cui, Shuang Chen, Yanzhi Wang, and 3 more authors

In 2016 21st Asia and South Pacific Design Automation Conference (ASP-DAC), Jan 2016

ISSN: 2153-697X

Abs DOI Bib

The heating, ventilation and air conditioning (HVAC) system accounts for half of the energy consumption of a typical building. Additionally, the need for HVAC changes over hours and days as does the electric energy price. Level of comfort of the building occupants is, however, a primary concern, which tends to overwrite pricing. Dynamic HVAC control under a dynamic energy pricing model while meeting an acceptable level of occupants’ comfort is thus critical to achieving energy efficiency in buildings in a sustainable manner. Finally, there is the possibility that the building is equipped with some renewable source of power such as solar panels mounted on the rooftop. The presence of a battery energy storage system in a target building would enable peak power shaving by adopting a suitable charge and discharge schedule for the battery, while simultaneously meeting building energy efficiency and user satisfaction. Achieving this goal requires detailed information (or predictions) about the amount of local power generation from the renewable source plus the power consumption load of the building. This paper addresses the coscheduling problem of HVAC control and battery management to achieve energy-efficient buildings, while also accounting for the degradation of the battery state-of-health during charging and discharging operations (which in turn determines the amortized cost of owning and utilizing a battery storage system)aa cč A time-of-use dynamic pricing scenario is assumed and various energy loss components are considered including power dissipation in the power conversion circuitry as well as the rate capacity effect in the battery. A global optimization framework targeting the entire billing cycle is presented and an adaptive co-scheduling algorithm is provided to dynamically update the optimal HVAC air flow control and the battery charging/discharging decision in each time slot during the billing cycle to mitigate the prediction error of unknown parameters. Experimental results show that the proposed algorithm achieves up to 15% in the total electric utility cost reduction compared with some baseline methods.
@inproceedings{cuiOptimalCoschedulingHVAC2016, title = {Optimal co-scheduling of {HVAC} control and battery management for energy-efficient buildings considering state-of-health degradation}, doi = {10.1109/ASPDAC.2016.7428105}, booktitle = {2016 21st {Asia} and {South} {Pacific} {Design} {Automation} {Conference} ({ASP}-{DAC})}, author = {Cui, Tiansong and Chen, Shuang and Wang, Yanzhi and Zhu, Qi and Nazarian, Shahin and Pedram, Massoud}, month = jan, year = {2016}, note = {ISSN: 2153-697X}, keywords = {Buildings, Control systems, Batteries, Degradation, Pricing, Power conversion}, pages = {775--780} }
Proactive Demand Participation of Smart Buildings in Smart Grid

Tianshu Wei, Qi Zhu, and Nanpeng Yu

IEEE Transactions on Computers, May 2016

Conference Name: IEEE Transactions on Computers

Abs DOI Bib

Buildings account for nearly 40 percent of the total energy consumption in the United States. As a critical step toward smart cities, it is essential to intelligently manage and coordinate the building operations to improve the efficiency and reliability of overall energy system. With the advent of smart meters and two-way communication systems, various energy consumptions from smart buildings can now be coordinated across the smart grid together with other energy loads and power plants. In this paper, we propose a comprehensive framework to integrate the operations of smart buildings into the energy scheduling of bulk power system through proactive building demand participation. This new scheme enables buildings to proactively express and communicate their energy consumption preferences to smart grid operators rather than passively receive and react to market signals and instructions such as time varying electricity prices. The proposed scheme is implemented in a simulation environment. The experiment results show that the proactive demand response scheme can achieve up to 10 percent system generation cost reduction and 20 percent building operation cost reduction compared with passive demand response scheme. The results also demonstrate that the system cost savings increase significantly with more flexible load installed and higher percentage of proactive customers participation level in the power network.
@article{weiProactiveDemandParticipation2016, title = {Proactive {Demand} {Participation} of {Smart} {Buildings} in {Smart} {Grid}}, volume = {65}, issn = {1557-9956}, doi = {10.1109/TC.2015.2495244}, number = {5}, journal = {IEEE Transactions on Computers}, author = {Wei, Tianshu and Zhu, Qi and Yu, Nanpeng}, month = may, year = {2016}, note = {Conference Name: IEEE Transactions on Computers}, keywords = {smart grid, Buildings, Energy consumption, Real-time systems, Batteries, Pricing, Load management, Load modeling, demand response, energy cost reduction, MPC control, proactive participation, Smart building}, pages = {1392--1406} }
ISCAS
Co-scheduling of flexible energy loads in building clusters

Tianshu Wei and Qi Zhu

In 2016 IEEE International Symposium on Circuits and Systems (ISCAS), May 2016

ISSN: 2379-447X

Abs DOI Bib

Buildings account for nearly 40% of energy consumption in the United States. To improve energy efficiency and reduce peak demand, intelligent building management systems can be developed to manage the energy consumption of flexible loads such as heating, ventilation, and air conditioning (HVAC) system an electric vehicle (EV) charging, as well as the usage of energy storage systems such as batteries. In the case where a building cluster is managed by the same institution, coordinating the energy consumption behavior across multiple buildings can provide further benefits in energy efficiency. In this paper, we first invest gate an integrated co-scheduling scheme that uses a joint formulation to optimize the control of HVAC systems, EV charging a d battery storage in multiple buildings for reducing the overall energy cost. Then, we further explore a more efficient heuristic scheme where the shared battery storage and EV charging demand are assigned to each building for separate building-level scheduling. Our experiments demonstrate the effectiveness of our co-scheduling scheme and separate-scheduling heuristic in reducing energy cost for building clusters.
@inproceedings{weiCoschedulingFlexibleEnergy2016, title = {Co-scheduling of flexible energy loads in building clusters}, doi = {10.1109/ISCAS.2016.7527401}, booktitle = {2016 {IEEE} {International} {Symposium} on {Circuits} and {Systems} ({ISCAS})}, author = {Wei, Tianshu and Zhu, Qi}, month = may, year = {2016}, note = {ISSN: 2379-447X}, keywords = {Mathematical model, Energy consumption, Batteries, Heating, Temperature, Windows}, pages = {958--961} }
Cross-Layer Codesign for Secure Cyber-Physical Systems

Bowen Zheng, Peng Deng, Rajasekhar Anguluri, and 2 more authors

IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems, May 2016

Conference Name: IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems

Abs DOI Bib

Security attacks may have disruptive consequences on cyber-physical systems, and lead to significant social and economic losses. Building secure cyber-physical systems is particularly challenging due to the variety of attack surfaces from the cyber and physical components, and often to limited computation and communication resources. In this paper, we propose a cross-layer design framework for resource-constrained cyber-physical systems. The framework combines control-theoretic methods at the functional layer and cybersecurity techniques at the embedded platform layer, and addresses security together with other design metrics such as control performance under resource and real-time constraints. We use the concept of interface variables to capture the interactions between control and platform layers, and quantitatively model the relation among system security, performance, and schedulability via interface variables. The general codesign framework is customized and refined to the automotive domain, and its effectiveness is demonstrated through an industrial case study and a set of synthetic examples.
@article{zhengCrossLayerCodesignSecure2016, title = {Cross-{Layer} {Codesign} for {Secure} {Cyber}-{Physical} {Systems}}, volume = {35}, issn = {1937-4151}, doi = {10.1109/TCAD.2016.2523937}, number = {5}, journal = {IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems}, author = {Zheng, Bowen and Deng, Peng and Anguluri, Rajasekhar and Zhu, Qi and Pasqualetti, Fabio}, month = may, year = {2016}, note = {Conference Name: IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems}, keywords = {Sensors, Automotive engineering, Cryptography, Control systems, Computational modeling, schedulability, cyber-physical systems, security, codesign, Codesign, control performance, cross-layer, Cyber-physical systems}, pages = {699--711} }
ISVLSI
Next Generation Automotive Architecture Modeling and Exploration for Autonomous Driving

Bowen Zheng, Hengyi Liang, Qi Zhu, and 2 more authors

In 2016 IEEE Computer Society Annual Symposium on VLSI (ISVLSI), Jul 2016

ISSN: 2159-3477

Abs DOI Bib

To support emerging applications in autonomous and semi-autonomous driving, next-generation automotive systems will be equipped with an increasing number of heterogeneous components (sensors, actuators and computation units connected through various buses), and have to process a high volume of data to percept the environment accurately and efficiently. Challenges for such systems include system integration, prediction, verification and validation. In this work, we propose an architecture modeling and exploration framework for evaluating various software and hardware architecture options. The framework will facilitate system integration and optimization, and enable validation of various design metrics such as timing, reliability, security and performance.
@inproceedings{zhengNextGenerationAutomotive2016, title = {Next {Generation} {Automotive} {Architecture} {Modeling} and {Exploration} for {Autonomous} {Driving}}, doi = {10.1109/ISVLSI.2016.126}, booktitle = {2016 {IEEE} {Computer} {Society} {Annual} {Symposium} on {VLSI} ({ISVLSI})}, author = {Zheng, Bowen and Liang, Hengyi and Zhu, Qi and Yu, Huafeng and Lin, Chung-Wei}, month = jul, year = {2016}, note = {ISSN: 2159-3477}, keywords = {Security, Computer architecture, Sensors, Automotive engineering, Autonomous automobiles, Software, Timing}, pages = {53--58} }
Fixed-Priority Dual-Rate Mixed-Criticality Systems Schedulability Analysis and Performance Optimization

Hang Su, Peng Deng, Dakai Zhu, and 1 more author

In , Aug 2016

Abs Bib

For mixed-criticality (MC) systems, recent studies show that it can be important to provide continuous (albeit degraded) services for low-critical (LC) tasks even in the high running mode. In this paper, focusing on dual-criticality systems, we study a mode-switch fixed-priority (MS-FP) scheduler for a set of dual-rate mixed-criticality (DR-MC) tasks, where each LC task can have a pair of small and large periods to represent its service requirements in the low (LO) and high (HI) running modes, respectively. Moreover, DR-MC tasks may adjust their priorities at the mode-switch point for better system schedulability. By extending the response time analysis (RTA) technique for MC systems, we first derive the schedulability conditions for a set of DR-MC tasks under the MS-FP scheduler with mode transition being considered. Then, we investigate how to select periods and priorities of DR-MC tasks to optimize their control performance and formulate it as a Non-Linear Optimization problem. We propose an efficient heuristic for a simplified optimization problem based on Branch & Bound Search Tree (BBST) technique. The effectiveness of the proposed heuristic and the MS-FP scheduler with DR-MC task model is illustrated through one case study with four tasks and compared against the Ipopt solutions.
@inproceedings{suFixedPriorityDualRateMixedCriticality2016, title = {Fixed-{Priority} {Dual}-{Rate} {Mixed}-{Criticality} {Systems} {Schedulability} {Analysis} and {Performance} {Optimization}}, author = {Su, Hang and Deng, Peng and Zhu, Dakai and Zhu, Qi}, month = aug, year = {2016} }

Automotive Cyber–Physical Systems: A Tutorial Introduction

Samarjit Chakraborty, Mohammad Abdullah Al Faruque, Wanli Chang, and 3 more authors

IEEE Design Test, Aug 2016

Conference Name: IEEE Design Test

Abs DOI Bib

This tutorial gives an introduction to novices in CPS and particularly highlights the basics of control theory with respect to automotive applications. The authors furthermore describe the “semantic gap” between control models and their implementation and conclude that a new CPS-oriented design approach is required.

@article{chakrabortyAutomotiveCyberPhysical2016,
  title = {Automotive {Cyber}–{Physical} {Systems}: {A} {Tutorial} {Introduction}},
  volume = {33},
  issn = {2168-2364},
  shorttitle = {Automotive {Cyber}–{Physical} {Systems}},
  doi = {10.1109/MDAT.2016.2573598},
  number = {4},
  journal = {IEEE Design Test},
  author = {Chakraborty, Samarjit and Al Faruque, Mohammad Abdullah and Chang, Wanli and Goswami, Dip and Wolf, Marilyn and Zhu, Qi},
  month = aug,
  year = {2016},
  note = {Conference Name: IEEE Design Test},
  keywords = {Computer architecture, Algorithm design and analysis, Automotive engineering, Program processors, Control theory, Software algorithms, Tutorials},
  pages = {92--108}
}

ICIP
Adaptive algorithm selection, with applications in pedestrian detection

Shu Zhang, Qi Zhu, and Amit Roy-Chowdhury

In 2016 IEEE International Conference on Image Processing (ICIP), Sep 2016

ISSN: 2381-8549

Abs DOI Bib

Computer vision algorithms are known to be extremely sensitive to the environmental conditions in which the data is captured, e.g., lighting conditions and target density. Tuning of parameters or choosing a completely new algorithm is often needed to achieve a certain performance level. In this paper, we focus on this problem and propose a framework to automatically choose the “best” algorithm-parameter combination (often referred to as the best algorithm for simplicity in this paper) for a certain input data. This necessitates developing a mechanism to switch among different algorithms and parameters as the nature of the input video changes. Specifically, our proposed algorithm calculates a similarity function between a test video segment and a training video segment. Similarity between training and test dataset indicates the same algorithm can be applied to both of them. We design a cost function with this similarity measure and a constraint on the number of switches. In the experiments, we apply our algorithm to the problem of pedestrian detection. We show how to adaptively select among 7 algorithm-parameter combinations and provide promising results on 3 publicly available datasets.
@inproceedings{zhangAdaptiveAlgorithmSelection2016, title = {Adaptive algorithm selection, with applications in pedestrian detection}, doi = {10.1109/ICIP.2016.7533064}, booktitle = {2016 {IEEE} {International} {Conference} on {Image} {Processing} ({ICIP})}, author = {Zhang, Shu and Zhu, Qi and Roy-Chowdhury, Amit}, month = sep, year = {2016}, note = {ISSN: 2381-8549}, keywords = {Training, adaptation, Algorithm design and analysis, Computer vision, Algorithm selection, Cost function, Detectors, Lighting, pedestrian detection, Switches}, pages = {3768--3772} }
ICCAD
CONVINCE: A cross-layer modeling, exploration and validation framework for next-generation connected vehicles

Bowen Zheng, Chung-Wei Lin, Huafeng Yu, and 2 more authors

In 2016 IEEE/ACM International Conference on Computer-Aided Design (ICCAD), Nov 2016

ISSN: 1558-2434

Abs DOI Bib

Next-generation autonomous and semi-autonomous vehicles will not only precept the environment with their own sensors, but also communicate with other vehicles and surrounding infrastructures for vehicle safety and transportation efficiency. The design, analysis and validation of various vehicle-to-vehicle (V2V) and vehicle-to-infrastructure (V2I) applications involve multiple layers, from V2V/V2I communication networks down to software and hardware of individual vehicles, and concern with stringent requirements on multiple metrics such as timing, security, reliability and fault tolerance. To cope with these challenges, we have been developing CONVINCE, a cross-layer modeling, exploration and validation framework for connected vehicles. The framework includes mathematical models, synthesis and validation algorithms, and a heterogeneous simulator for inter-vehicle communications and intra-vehicle software and hardware in a holistic environment. It explores various design options with respect to constraints and objectives on system safety, security, reliability, cost, etc. A V2V application is used in the case study to demonstrate the effectiveness of the proposed framework.
@inproceedings{zhengCONVINCECrosslayerModeling2016, title = {{CONVINCE}: {A} cross-layer modeling, exploration and validation framework for next-generation connected vehicles}, shorttitle = {{CONVINCE}}, doi = {10.1145/2966986.2980078}, booktitle = {2016 {IEEE}/{ACM} {International} {Conference} on {Computer}-{Aided} {Design} ({ICCAD})}, author = {Zheng, Bowen and Lin, Chung-Wei and Yu, Huafeng and Liang, Hengyi and Zhu, Qi}, month = nov, year = {2016}, note = {ISSN: 1558-2434}, keywords = {Security, Vehicles, Mathematical model, Computational modeling, Software, Timing, Analytical models}, pages = {1--8} }
IGSC
Co-scheduling of datacenter and HVAC loads in mixed-use buildings

Tianshu Wei, Mohammad Atiqul Islam, Shaolei Ren, and 1 more author

In 2016 Seventh International Green and Sustainable Computing Conference (IGSC), Nov 2016

Abs DOI Bib

The majority of datacenters are within mixed-use facilities, where they often share some common infrastructures and energy supplies with other operations (e.g., non-IT offices and labs). In such mixed-use buildings, two major energy loads are datacenter IT equipment and HVAC (heating, ventilating, and air conditioning) system. The HVAC demand comes from both datacenter rooms and other non-IT rooms. To effectively lower peak demand and reduce energy cost for mixed-use buildings, it is important to leverage the scheduling flexibility from both the HVAC system and the delay-tolerant datacenter workload in a collaborative fashion. In this work, we model the major physical and cyber components of mixed-use buildings, and propose a model predictive control (MPC) formulation to co-schedule datacenter and HVAC loads, with consideration of solar energy and battery storage. The MPC formulation minimizes building energy cost while satisfying various requirements on room temperature, ventilation, and datacenter workload deadlines. Compared with separate scheduling strategy, our approach significantly reduces peak demand and overall energy cost, and provides better leverage of renewable energy supply. Furthermore, we demonstrate that our formulation is also effective in reducing carbon footprint, and balancing its trade-off with energy cost.
@inproceedings{weiCoschedulingDatacenterHVAC2016, title = {Co-scheduling of datacenter and {HVAC} loads in mixed-use buildings}, doi = {10.1109/IGCC.2016.7892609}, booktitle = {2016 {Seventh} {International} {Green} and {Sustainable} {Computing} {Conference} ({IGSC})}, author = {Wei, Tianshu and Islam, Mohammad Atiqul and Ren, Shaolei and Zhu, Qi}, month = nov, year = {2016}, keywords = {Buildings, Mathematical model, Atmospheric modeling, Load modeling, Ventilation, Heating systems}, pages = {1--8} }
An Efficient Control-Driven Period Optimization Algorithm for Distributed Real-Time Systems

Peng Deng, Qi Zhu, Abhijit Davare, and 3 more authors

IEEE Transactions on Computers, Dec 2016

Conference Name: IEEE Transactions on Computers

Abs DOI Bib

The sampling periods of real-time embedded control functions have a significant impact on control performance and system schedulability. Exploring period assignment for optimizing control performance while meeting schedulability constraints is very challenging, in particular for distributed systems where control loops share a network of computation and communication resources. In this work, we propose an efficient approach that approximates the performance of each control loop in the system with a piecewise linear function of its sampling period and end-to-end delay, and then optimizes the periods of tasks and messages by exploring the linear partitions of the approximated functions and solving a series of geometric programming (GP) formulations. Experiments on sample control models, an automotive industrial case study and a set of synthetic examples demonstrate the effectiveness and efficiency of our approach.
@article{dengEfficientControlDrivenPeriod2016, title = {An {Efficient} {Control}-{Driven} {Period} {Optimization} {Algorithm} for {Distributed} {Real}-{Time} {Systems}}, volume = {65}, issn = {1557-9956}, doi = {10.1109/TC.2016.2557322}, number = {12}, journal = {IEEE Transactions on Computers}, author = {Deng, Peng and Zhu, Qi and Davare, Abhijit and Mourikis, Anastasios and Liu, Xue and Natale, Marco Di}, month = dec, year = {2016}, note = {Conference Name: IEEE Transactions on Computers}, keywords = {Control systems, Optimization, Real-time systems, Partitioning algorithms, Programming, geometric programming, Geometric programming, period optimization, Real-time embedded control system, Search methods}, pages = {3552--3566} }

2015

Security-Aware Modeling and Efficient Mapping for CAN-Based Real-Time Distributed Automotive Systems

Chung-Wei Lin, Qi Zhu, and Alberto Sangiovanni-Vincentelli

IEEE Embedded Systems Letters, Mar 2015

Conference Name: IEEE Embedded Systems Letters

Abs DOI Bib

Security has become a critical issue for automotive electronic systems. To protect against attacks, security mechanisms have to be applied, but the overhead of those mechanisms may impede system performance and cause violations of design constraints. To remedy this problem, we proposed an integrated mixed integer linear programming (MILP) formulation that is the first to address both security and safety constraints during system mapping for controller area network (CAN) based systems . However, its signal-based security constraints do not fully reflect real security requirements, and its objective function is to minimize functional path latencies rather than minimize security risk. Furthermore, its MILP-based approach has high computation complexity. In this work, we present a new formulation that defines path-based security constraints and minimizes security risk directly, and propose a new heuristic algorithm to solve the formulation efficiently. Experiments on an industrial example show that our new algorithm achieves comparable solution quality as the MILP-based approach with much better efficiency.
@article{linSecurityAwareModelingEfficient2015, title = {Security-{Aware} {Modeling} and {Efficient} {Mapping} for {CAN}-{Based} {Real}-{Time} {Distributed} {Automotive} {Systems}}, volume = {7}, issn = {1943-0671}, doi = {10.1109/LES.2014.2354011}, number = {1}, journal = {IEEE Embedded Systems Letters}, author = {Lin, Chung-Wei and Zhu, Qi and Sangiovanni-Vincentelli, Alberto}, month = mar, year = {2015}, note = {Conference Name: IEEE Embedded Systems Letters}, keywords = {Security, Heuristic algorithms, Resource management, Automotive engineering, Runtime, design space exploration, Equations, Automotive systems, cyber-physical systems, embedded systems, Linear programming, security}, pages = {11--14} }
Design and Operation of Secure Cyber-Physical Systems

Fabio Pasqualetti and Qi Zhu

IEEE Embedded Systems Letters, Mar 2015

Conference Name: IEEE Embedded Systems Letters

Abs DOI Bib

This letter proposes a holistic framework for the design and operation of secure and reliable resource-constrained cyber-physical systems. The proposed framework combines control-theoretic methods, information security notions and computational models to characterize tradeoffs among different design and operation objectives. We quantify the intricate relation among control performance, system security and platform schedulability through a minimal set of interface variables. We argue that security mechanisms and control algorithms need to be codesigned and comanaged with the embedded platform, so as to avoid the design of algorithms that are too expensive to implement on the embedded platform, or significantly impede design objectives such as performance and timing robustness.
@article{pasqualettiDesignOperationSecure2015, title = {Design and {Operation} of {Secure} {Cyber}-{Physical} {Systems}}, volume = {7}, issn = {1943-0671}, doi = {10.1109/LES.2014.2367100}, number = {1}, journal = {IEEE Embedded Systems Letters}, author = {Pasqualetti, Fabio and Zhu, Qi}, month = mar, year = {2015}, note = {Conference Name: IEEE Embedded Systems Letters}, keywords = {Algorithm design and analysis, Sensors, Control systems, Real-time systems, Automotive systems, embedded systems, security, control theory, cyber-physical system, Encryption}, pages = {3--6} }
Task placement and selection of data consistency mechanisms for real-time multicore applications

Zaid Al-bayati, Youcheng Sun, Haibo Zeng, and 3 more authors

In 21st IEEE Real-Time and Embedded Technology and Applications Symposium, Apr 2015

ISSN: 1545-3421

Abs DOI Bib

Multicores are today used in automotive, controls and avionics systems supporting real-time functionality. When real-time tasks allocated on different cores cooperate through the use of shared communication resources, they need to be protected by mechanisms that guarantee access in a mutual exclusive way with bounded worst-case blocking time. Lock-based mechanisms such as MPCP and MSRP have been developed to fulfill this demand, and research papers are today tackling the problem of finding the optimal task placement in multicores while trying to meet the deadlines against blocking times. In this paper, we propose a resource-aware task allocation algorithm for systems that use MSRP to protect shared resources. Furthermore, we leverage the additional opportunity provided by wait-free methods as an alternative data consistency mechanism for the case that the shared resource is communication or state memory. An algorithm that performs both task allocation and data consistency mechanism (MSRP or wait-free) selection is proposed. The selective use of wait-free methods can significantly extend the range of schedulable systems at the cost of memory.
@inproceedings{al-bayatiTaskPlacementSelection2015, title = {Task placement and selection of data consistency mechanisms for real-time multicore applications}, doi = {10.1109/RTAS.2015.7108440}, booktitle = {21st {IEEE} {Real}-{Time} and {Embedded} {Technology} and {Applications} {Symposium}}, author = {Al-bayati, Zaid and Sun, Youcheng and Zeng, Haibo and Di Natale, Marco and Zhu, Qi and Meyer, Brett}, month = apr, year = {2015}, note = {ISSN: 1545-3421}, keywords = {Resource management, Protocols, Real-time systems, Partitioning algorithms, Time factors, Hardware, Multicore processing}, pages = {172--181} }
A model-based synthesis flow for automotive CPS

Peng Deng, Fabio Cremona, Qi Zhu, and 2 more authors

In Proceedings of the ACM/IEEE Sixth International Conference on Cyber-Physical Systems, Apr 2015

Abs DOI Bib HTML

Synchronous reactive models are used by automotive suppliers to develop functionality delivered as AUTOSAR components to system integrators (OEMs). Integrators must then generate a task implementation from runnables in AUTOSAR components and deploy tasks onto CPU cores, while preserving timing and resource constraints. In this work, we propose an integrated synthesis flow that addresses both sides of the supply chain. On the supplier side, from synchronous models, we generate AUTOSAR runnables that promote reuse and ease the job of finding schedulable implementations. On the integrator side, we find the mapping of runnables onto tasks and allocation of tasks on cores that satisfy the timing constraints and are memory efficient.
@inproceedings{dengModelbasedSynthesisFlow2015, address = {New York, NY, USA}, series = {{ICCPS} '15}, title = {A model-based synthesis flow for automotive {CPS}}, isbn = {978-1-4503-3455-6}, url = {https://doi.org/10.1145/2735960.2735972}, doi = {10.1145/2735960.2735972}, urldate = {2022-05-05}, booktitle = {Proceedings of the {ACM}/{IEEE} {Sixth} {International} {Conference} on {Cyber}-{Physical} {Systems}}, publisher = {Association for Computing Machinery}, author = {Deng, Peng and Cremona, Fabio and Zhu, Qi and Di Natale, Marco and Zeng, Haibo}, month = apr, year = {2015}, pages = {198--207} }
PeerWave: Exploiting Wavefront Parallelism on GPUs with Peer-SM Synchronization

Mehmet E. Belviranli, Peng Deng, Laxmi N. Bhuyan, and 2 more authors

In Proceedings of the 29th ACM on International Conference on Supercomputing, Jun 2015

Abs DOI Bib HTML

Nested loops with regular iteration dependencies span a large class of applications ranging from string matching to linear system solvers. Wavefront parallelism is a well-known technique to enable concurrent processing of such applications and is widely being used on GPUs to benefit from their massively parallel computing capabilities. Wavefront parallelism on GPUs uses global barriers between processing of tiles to enforce data dependencies. However, such diagonal-wide synchronization causes load imbalance by forcing SMs to wait for the completion of the SM with longest computation. Moreover, diagonal processing causes loss of locality due to elements that border adjacent tiles. In this paper, we propose PeerWave, an alternative GPU wavefront parallelization technique that improves inter-SM load balance by using peer-wise synchronization between SMs. and eliminating global synchronization. Our approach also increases GPU L2 cache locality through row allocation of tiles to the SMs. We further improve PeerWave performance by using flexible hyper-tiles that reduce inter-SM wait time while maximizing intra-SM utilization. We develop an analytical model for determining the optimal tile size. Finally, we present a run-time and a CUDA based API to allow users to easily implement their applications using PeerWave. We evaluate PeerWave on the NVIDIA K40c GPU using 6 different applications and achieve speedups of up to 2X compared to the most recent hyperplane transformation based GPU implementation.
@inproceedings{belviranliPeerWaveExploitingWavefront2015, address = {New York, NY, USA}, series = {{ICS} '15}, title = {{PeerWave}: {Exploiting} {Wavefront} {Parallelism} on {GPUs} with {Peer}-{SM} {Synchronization}}, isbn = {978-1-4503-3559-1}, shorttitle = {{PeerWave}}, url = {https://doi.org/10.1145/2751205.2751243}, doi = {10.1145/2751205.2751243}, urldate = {2022-05-05}, booktitle = {Proceedings of the 29th {ACM} on {International} {Conference} on {Supercomputing}}, publisher = {Association for Computing Machinery}, author = {Belviranli, Mehmet E. and Deng, Peng and Bhuyan, Laxmi N. and Gupta, Rajiv and Zhu, Qi}, month = jun, year = {2015}, keywords = {decentralized synchronization, gp-gpu computing, wavefront parallelism}, pages = {25--35} }
DAC
Optimal control of PEVs for energy cost minimization and frequency regulation in the smart grid accounting for battery state-of-health degradation

Tiansong Cui, Yanzhi Wang, Shuang Chen, and 3 more authors

In 2015 52nd ACM/EDAC/IEEE Design Automation Conference (DAC), Jun 2015

ISSN: 0738-100X

Abs DOI Bib

Plug-in electric vehicles (PEVs) are considered the key to reducing the fossil fuel consumption and an important part of the smart grid. The plug-in electric vehicle-to-grid (V2G) technology in the smart grid infrastructure enables energy flow from PEV batteries to the power grid so that the grid stability is enhanced and the peak power demand is shaped. PEV owners will also benefit from V2G technology as they will be able to reduce energy cost through proper PEV charging and discharging scheduling. Moreover, power regulation service (RS) reserves have been playing an increasingly important role in modern power markets. It has been shown that by providing RS reserves, the power grid achieves a better match between energy supply and demand in presence of volatile and intermittent renewable energy generation. This paper addresses the problem of PEV charging under dynamic energy pricing, properly taking into account the degradation of battery state-of-health (SoH) during V2G operations as well as RS provisioning. An overall optimization throughout the whole parking period is proposed for the PEV and an adaptive control framework is presented to dynamically update the optimal charging/discharging decision at each time slot to mitigate the effect of RS tracking error. Experimental results show that the proposed optimal PEV charging algorithm minimizes the combination of electricity cost and battery aging cost in the RS provisioning power market.
@inproceedings{cuiOptimalControlPEVs2015, title = {Optimal control of {PEVs} for energy cost minimization and frequency regulation in the smart grid accounting for battery state-of-health degradation}, doi = {10.1145/2744769.2744882}, booktitle = {2015 52nd {ACM}/{EDAC}/{IEEE} {Design} {Automation} {Conference} ({DAC})}, author = {Cui, Tiansong and Wang, Yanzhi and Chen, Shuang and Zhu, Qi and Nazarian, Shahin and Pedram, Massoud}, month = jun, year = {2015}, note = {ISSN: 0738-100X}, keywords = {Power demand, Batteries, System-on-chip, Aging, Degradation, Power markets, Pricing}, pages = {1--6} }
DAC
Design and verification for transportation system security

Bowen Zheng, Wenchao Li, Peng Deng, and 3 more authors

In 2015 52nd ACM/EDAC/IEEE Design Automation Conference (DAC), Jun 2015

ISSN: 0738-100X

Abs DOI Bib

Cyber-security has emerged as a pressing issue for transportation systems. Studies have shown that attackers can attack modern vehicles from a variety of interfaces and gain access to the most safety-critical components. Such threats become even broader and more challenging with the emergence of vehicle-to-vehicle (V2V) and vehicle-to-infrastructure (V2I) communication technologies. Addressing the security issues in transportation systems requires comprehensive approaches that encompass considerations of security mechanisms, safety properties, resource constraints, and other related system metrics. In this work, we propose an integrated framework that combines hybrid modeling, formal verification, and automated synthesis techniques for analyzing the security and safety of transportation systems and carrying out design space exploration of both in-vehicle electronic control systems and vehicle-to-vehicle communications. We demonstrate the ideas of our framework through a case study of cooperative adaptive cruise control.
@inproceedings{zhengDesignVerificationTransportation2015, title = {Design and verification for transportation system security}, doi = {10.1145/2744769.2747920}, booktitle = {2015 52nd {ACM}/{EDAC}/{IEEE} {Design} {Automation} {Conference} ({DAC})}, author = {Zheng, Bowen and Li, Wenchao and Deng, Peng and Gérardy, Léonard and Zhu, Qi and Shankar, Natarajan}, month = jun, year = {2015}, note = {ISSN: 0738-100X}, keywords = {Security, Safety, Sensors, Vehicles, Delays}, pages = {1--6} }
CASE
From passive demand response to proactive demand participation

Nanpeng Yu, Tianshu Wei, and Qi Zhu

In 2015 IEEE International Conference on Automation Science and Engineering (CASE), Aug 2015

ISSN: 2161-8089

Abs DOI Bib

Limited progress has been made in the past few years in increasing demand response participation rate in the United States. The structural rigidity of existing price-based and incentive-based demand response programs results in inadequate and inefficient utilization of demand flexibility in electricity market operations. In this paper, an innovative proactive demand participation scheme is developed. This new scheme enables customers to actively express and communicate their consumption preferences to market operators rather than passively receive and react to time varying electricity prices and demand reduction signals. A novel framework for integrated wholesale and retail market operations with proactive demand participation and customer aggregation is proposed. The proactive demand response scheme is implemented in a simulation environment. The simulation results show that the proactive demand participation scheme is superior to the passive demand response approach. The proactive demand participation approach not only increases overall market efficiency but also reduces price volatility.
@inproceedings{yuPassiveDemandResponse2015, title = {From passive demand response to proactive demand participation}, doi = {10.1109/CoASE.2015.7294278}, booktitle = {2015 {IEEE} {International} {Conference} on {Automation} {Science} and {Engineering} ({CASE})}, author = {Yu, Nanpeng and Wei, Tianshu and Zhu, Qi}, month = aug, year = {2015}, note = {ISSN: 2161-8089}, keywords = {Buildings, Mathematical model, Real-time systems, Batteries, Building Aggregation, Demand Response, Generators, Integrated Market, Load management, Load modeling, Model Predictive Control, Proactive Demand Participation}, pages = {1300--1306} }
CODES+ISSS
Analysis and optimization of soft error tolerance strategies for real-time systems

Bowen Zheng, Yue Gao, Qi Zhu, and 1 more author

In 2015 International Conference on Hardware/Software Codesign and System Synthesis (CODES+ISSS), Oct 2015

Abs DOI Bib

The safety of real-time embedded systems relies on both functional and timing correctness. On the timing side, realtime constraints are set on task executions, and missing them may lead to system failure. On the functional side, soft errors have become a major concern. Various soft error tolerance strategies are proposed for soft error detection and recovery, however they may introduce significant computation overhead and cause timing violations. In this work, we address the two aspects in an integrated framework, and propose a set of formulations to quantitatively model the impact of soft error detection and recovery mechanisms on real-time constraints. The formulations facilitate designers to analyze system feasibility under fault tolerance requirements and compare various architecture platforms. They may also help select the appropriate error tolerance mechanisms for software tasks, together with exploring task scheduling and allocation on representative single-core, multicore and distributed platforms, to maximize error coverage while meeting real-time constraints. Experiments on an industrial case study and synthetic examples demonstrate the effectiveness of our approach.
@inproceedings{zhengAnalysisOptimizationSoft2015, title = {Analysis and optimization of soft error tolerance strategies for real-time systems}, doi = {10.1109/CODESISSS.2015.7331368}, booktitle = {2015 {International} {Conference} on {Hardware}/{Software} {Codesign} and {System} {Synthesis} ({CODES}+{ISSS})}, author = {Zheng, Bowen and Gao, Yue and Zhu, Qi and Gupta, Sandeep}, month = oct, year = {2015}, keywords = {Resource management, Fault tolerance, Fault tolerant systems, Timing, Real-time systems, Time factors, Multicore processing}, pages = {55--64} }
IGSC
Proactive demand participation of heterogeneous flexible loads in smart grid

Tianshu Wei and Qi Zhu

In 2015 Sixth International Green and Sustainable Computing Conference (IGSC), Dec 2015

Abs DOI Bib

There are a variety of flexible energy demand loads in modern buildings. In particular, heating, ventilation and air conditioning (HVAC) system accounts for around 50% of buildings’ total energy consumption and is a major source for energy scheduling flexibility. Electric vehicle (EV) charging is an emerging load in many buildings, and may also provides scheduling flexibility depending on the charging requirements. In addition, the usage of energy storage systems such as batteries further increases the flexibility in scheduling building energy demands. At the power grid level, it is important to leverage such flexibility for improving energy efficiency. To this end, we have proposed a proactive demand response scheme that enables building customers to actively participate in the electricity market clearing process. Compared to the conventional passive demand response strategy, the proactive scheme is shown to be more effective to take advantage of buildings’ flexible load demand and enhance the energy efficiency of the overall power system. In this paper, the scheduling of various flexible building energy loads such as HVAC control and EV charging, as well as the usage of battery storage system, are jointly modeled and optimized in the proactive demand participation framework. We conduct experiments on the IEEE 5-bus system to evaluate the impact of heterogeneous flexible loads on the energy efficiency of the proactive scheme.
@inproceedings{weiProactiveDemandParticipation2015, title = {Proactive demand participation of heterogeneous flexible loads in smart grid}, doi = {10.1109/IGCC.2015.7393722}, booktitle = {2015 {Sixth} {International} {Green} and {Sustainable} {Computing} {Conference} ({IGSC})}, author = {Wei, Tianshu and Zhu, Qi}, month = dec, year = {2015}, keywords = {Buildings, Energy consumption, Atmospheric modeling, Batteries, Load management, Heating, Ventilation}, pages = {1--2} }
ICCAD
Security analysis of proactive participation of smart buildings in smart grid

Tianshu Wei, Bowen Zheng, Qi Zhu, and 1 more author

In 2015 IEEE/ACM International Conference on Computer-Aided Design (ICCAD), Nov 2015

Abs DOI Bib

Demand response (DR) is an effective mechanism in improving power system efficiency and reducing energy cost for customers. However, DR processes might be vulnerable to cyber attacks from the usage of advanced metering infrastructure and wide-area network to exchange information. In this paper, we study potential attacks for a proactive demand participation scheme we recently proposed and for a conventional passive demand response scheme, particularly focusing on guideline price manipulation attacks. Our experiment results demonstrate that 1) guideline price manipulations may significantly lower the attacker’s own electricity consumption cost while increasing other customers’ cost, for both proactive and passive schemes; 2) such impact is less severe in the proactive scheme, i.e., the proactive demand participation scheme is more robust with respect to guideline price manipulation than the conventional DR.
@inproceedings{weiSecurityAnalysisProactive2015, title = {Security analysis of proactive participation of smart buildings in smart grid}, doi = {10.1109/ICCAD.2015.7372606}, booktitle = {2015 {IEEE}/{ACM} {International} {Conference} on {Computer}-{Aided} {Design} ({ICCAD})}, author = {Wei, Tianshu and Zheng, Bowen and Zhu, Qi and Hu, Shiyan}, month = nov, year = {2015}, keywords = {Buildings, Energy consumption, Load management, Electricity supply industry, Guidelines, Power generation, Substations}, pages = {465--472} }
Security-Aware Design Methodology and Optimization for Automotive Systems

Chung-Wei Lin, Bowen Zheng, Qi Zhu, and 1 more author

ACM Transactions on Design Automation of Electronic Systems, Dec 2015

Abs DOI Bib HTML

In this article, we address both security and safety requirements and solve security-aware design problems for the controller area network (CAN) protocol and time division multiple access (TDMA)-based protocols. To provide insights and guidelines for other similar security problems with limited resources and strict timing constraints, we propose a general security-aware design methodology to address security with other design constraints in a holistic framework and optimize design objectives. The security-aware design methodology is further applied to solve a security-aware design problem for vehicle-to-vehicle (V2V) communications with dedicated short-range communication (DSRC) technology. Experimental results demonstrate the effectiveness of our approaches in system design without violating design constraints and indicate that it is necessary to consider security together with other metrics during design stages.
@article{linSecurityAwareDesignMethodology2015, title = {Security-{Aware} {Design} {Methodology} and {Optimization} for {Automotive} {Systems}}, volume = {21}, issn = {1084-4309}, url = {https://doi.org/10.1145/2803174}, doi = {10.1145/2803174}, number = {1}, urldate = {2022-05-06}, journal = {ACM Transactions on Design Automation of Electronic Systems}, author = {Lin, Chung-Wei and Zheng, Bowen and Zhu, Qi and Sangiovanni-Vincentelli, Alberto}, month = dec, year = {2015}, keywords = {automotive systems, cyber-physical systems, controller area network, dedicated short-range communication, mapping, Methodology, time division multiple access, time-triggered ethernet}, pages = {18:1--18:26} }

2014

Minimizing Stack and Communication Memory Usage in Real-Time Embedded Applications

Haibo Zeng, Marco Di Natale, and Qi Zhu

ACM Transactions on Embedded Computing Systems, Jul 2014

Abs DOI Bib HTML

In the development of real-time embedded applications, especially those on systems-on-chip, an efficient use of RAM memory is as important as the effective scheduling of the computation resources. The protection of communication and state variables accessed by concurrent tasks must provide real-time schedulability guarantees while using the least amount of memory. Several schemes, including preemption thresholds, have been developed to improve schedulability and save stack space by selectively disabling preemption. However, the design synthesis problem is still open. In this article, we target the assignment of the scheduling parameters to minimize memory usage for systems of practical interest, including designs compliant with automotive standards. We propose algorithms either proven optimal or shown to improve on randomized optimization methods like simulated annealing.
@article{zengMinimizingStackCommunication2014, title = {Minimizing {Stack} and {Communication} {Memory} {Usage} in {Real}-{Time} {Embedded} {Applications}}, volume = {13}, issn = {1539-9087}, url = {https://doi.org/10.1145/2632160}, doi = {10.1145/2632160}, number = {5s}, urldate = {2022-05-06}, journal = {ACM Transactions on Embedded Computing Systems}, author = {Zeng, Haibo and Natale, Marco Di and Zhu, Qi}, month = jul, year = {2014}, keywords = {data synchronization mechanism, memory usage, Preemption threshold scheduling, stack requirement}, pages = {149:1--149:25} }
Optimized implementation of synchronous models on industrial LTTA systems

Marco Di Natale, Qi Zhu, Alberto Sangiovanni-Vincentelli, and 1 more author

Journal of Systems Architecture, Apr 2014

Abs DOI Bib HTML

Synchronous models are used to specify embedded systems functions in a clear and unambiguous way and allow verification of properties using formal methods. The implementation of a synchronous specification on a distributed architecture must preserve the model semantics to retain the verification results. Globally synchronized time-triggered architectures offer the simplest implementation path, but can be inefficient or simply unavailable. In past work, we defined a mapping of synchronous models on a general class of distributed asynchronous architectures, for which the only requirement is a lower bound on the rate of activation of tasks. In this paper, we set tighter requirements on task execution rates, and we include a realistic modeling of communication delays, task scheduling delays and schedulability conditions, discussing the timing characteristics of an implementation on a system with a Controller Area Network (CAN). Next, the semantics preservation conditions are formulated as constraints in an architecture optimization problem that defines a feasible task model with respect to timing constraints. An automotive case study shows the applicability of the approach and provides insight on the software design elements that are critical for a feasible implementation.
@article{dinataleOptimizedImplementationSynchronous2014, title = {Optimized implementation of synchronous models on industrial {LTTA} systems}, volume = {60}, issn = {1383-7621}, url = {https://www.sciencedirect.com/science/article/pii/S1383762114000198}, doi = {10.1016/j.sysarc.2014.01.003}, language = {en}, number = {4}, urldate = {2022-05-06}, journal = {Journal of Systems Architecture}, author = {Di Natale, Marco and Zhu, Qi and Sangiovanni-Vincentelli, Alberto and Tripakis, Stavros}, month = apr, year = {2014}, keywords = {Code generation, Distributed systems, Embedded systems, Loosely time-triggered architecture, Semantics-preserving implementation, Synchronous models}, pages = {315--328} }
DATE
MSim: A general cycle accurate simulation platform for memcomputing studies

Chun Zhang, Peng Deng, Hui Geng, and 4 more authors

In 2014 Design, Automation Test in Europe Conference Exhibition (DATE), Mar 2014

ISSN: 1558-1101

Abs DOI Bib

The lack of accurate yet open to public simulation infrastructure has puzzled researchers in the memcomputing area for sometime. In this paper, we propose for the first time a full tool chain called MSim that supports the cycle-accurate microarchitecture level simulation for memcomputing studies. With MSim, the performance gains of utilizing memcomputing for arbitrary applications on user configurable computer system architectures can be evaluated in high accuracy. In addition, MSim provides flexible interfaces with pervasive object-oriented design, which makes it well-suited as a good base platform for researchers to explore new memcomputing technologies.
@inproceedings{zhangMSimGeneralCycle2014, title = {{MSim}: {A} general cycle accurate simulation platform for memcomputing studies}, shorttitle = {{MSim}}, doi = {10.7873/DATE.2014.278}, booktitle = {2014 {Design}, {Automation} {Test} in {Europe} {Conference} {Exhibition} ({DATE})}, author = {Zhang, Chun and Deng, Peng and Geng, Hui and Liu, Jianming and Zhu, Qi and Xiong, Jinjun and Shi, Yiyu}, month = mar, year = {2014}, note = {ISSN: 1558-1101}, keywords = {Computational modeling, Computers, Engines, Microarchitecture, Object oriented modeling, Table lookup}, pages = {1--5} }
Design synthesis and optimization for automotive embedded systems

Qi Zhu and Peng Deng

In Proceedings of the 2014 on International symposium on physical design, Mar 2014

Abs DOI Bib HTML

Embedded software and electronics are major contributors of values in vehicles, and play a dominant role in vehicle innovations. The design of automotive embedded systems has become more and more challenging, with the rapid increase of system complexity and more requirements on various design objectives. Methodologies such as model-based design are being adopted to improve design quality and productivity through the usage of functional models. However, there is still a significant lack of design automation tools, in particular synthesis and optimization tools, that can turn complex functional specifications to correct and optimal software implementations on distributed embedded platforms. In this paper, we discuss some of the major technical challenges and the problems to be solved in automotive embedded systems design, especially for the synthesis and optimization of embedded software.
@inproceedings{zhuDesignSynthesisOptimization2014, address = {New York, NY, USA}, series = {{ISPD} '14}, title = {Design synthesis and optimization for automotive embedded systems}, isbn = {978-1-4503-2592-9}, url = {https://doi.org/10.1145/2560519.2565873}, doi = {10.1145/2560519.2565873}, urldate = {2022-05-05}, booktitle = {Proceedings of the 2014 on {International} symposium on physical design}, publisher = {Association for Computing Machinery}, author = {Zhu, Qi and Deng, Peng}, month = mar, year = {2014}, keywords = {automotive embedded systems, design automation, software synthesis and optimization}, pages = {141--148} }
ICCPS
WiP abstract: An efficient control-driven period optimization algorithm for distributed real-time systems

Peng Deng, Anastasios Mourikis, Qi Zhu, and 3 more authors

In 2014 ACM/IEEE International Conference on Cyber-Physical Systems (ICCPS), Apr 2014

Abs DOI Bib

The sampling periods of real-time embedded control functions have a significant impact on control performance and system schedulability. Exploring period assignment for optimizing control performance while meeting schedulability constraints is very challenging, in particular for distributed systems where control loops share computation and communication resources. We propose an efficient approach that approximates the performance of each control loop in the system with a piece-wise linear function of its sampling period and end-to-end delay, and then optimizes the periods of tasks and messages by exploring the linear partitions of the approximated functions and solving a series of geometric programming (GP) formulations.
@inproceedings{dengWiPAbstractEfficient2014, title = {{WiP} abstract: {An} efficient control-driven period optimization algorithm for distributed real-time systems}, shorttitle = {{WiP} abstract}, doi = {10.1109/ICCPS.2014.6843728}, booktitle = {2014 {ACM}/{IEEE} {International} {Conference} on {Cyber}-{Physical} {Systems} ({ICCPS})}, author = {Deng, Peng and Mourikis, Anastasios and Zhu, Qi and Liu, Xue and Davare, Abhijit and Di Natale, Marco}, month = apr, year = {2014}, keywords = {Control systems, Optimization, Real-time systems, Abstracts, Approximation methods, Piecewise linear approximation, Programming}, pages = {215--215} }
DAC
Battery management and application for energy-efficient buildings

Tianshu Wei, Taeyoung Kim, Sangyoung Park, and 5 more authors

In 2014 51st ACM/EDAC/IEEE Design Automation Conference (DAC), Jun 2014

ISSN: 0738-100X

Abs DOI Bib

As the building stock consumes 40% of the U.S. primary energy consumption, it is critically important to improve building energy efficiency. This involves reducing the total energy consumption of buildings, reducing the peak energy demand, and leveraging renewable energy sources, etc. To achieve such goals, hybrid energy supply has becoming popular, where multiple energy sources such as grid electricity, on-site fuel cell generators, solar, wind, and battery storage are scheduled together to improve energy efficiency. In this work, we focus on the application and management of battery storage for energy-efficient buildings. We will first introduce a system-level approach to co-schedule the usage of battery storage (in addition to grid electricity) with the control of building HVAC (heating, ventilation, and air conditioning) system, to reduce the total building energy cost, including the electricity consumption charge, the peak demand charge, and the battery cost. Then, in a separate formulation, we will introduce another system-level study to reduce the energy cost of EV charging and other fixed building energy load through the usage of battery storage and solar PV. Finally, we will present an ARM processor based programmable embedded battery management system (BMS), which monitors battery status, controls charging and discharging at the circuit level, and provides battery protection. The system also works with off-the-shelf battery management IC (Texas Instrument BMS sensor IC) from industry. Comparing to conventional BMS, this software module based BMS is a more suitable solution for energy efficient buildings due to its high flexibility, scalability, and reusability. We will introduce an industrial building testbed with battery storage and solar PV at the University of California, Riverside, and present initial field tests and simulation results for above approaches.
@inproceedings{weiBatteryManagementApplication2014, title = {Battery management and application for energy-efficient buildings}, doi = {10.1145/2593069.2596670}, booktitle = {2014 51st {ACM}/{EDAC}/{IEEE} {Design} {Automation} {Conference} ({DAC})}, author = {Wei, Tianshu and Kim, Taeyoung and Park, Sangyoung and Zhu, Qi and Tan, Sheldon X.-D. and Chang, Naehyuck and Ula, Sadrul and Maasoumy, Mehdi}, month = jun, year = {2014}, note = {ISSN: 0738-100X}, keywords = {Power grids, Buildings, Batteries, Battery charge measurement, Electricity, Monitoring, System-on-chip}, pages = {1--6} }
SIES
Task synthesis for latency-sensitive synchronous block diagram

Peng Deng, Qi Zhu, Marco Di Natale, and 1 more author

In Proceedings of the 9th IEEE International Symposium on Industrial Embedded Systems (SIES 2014), Jun 2014

ISSN: 2150-3117

Abs DOI Bib

Synchronous block diagrams (SBDs) are commonly used in model-based design tools such as Simulink to capture the system behavior. In the multitask software implementation of SBDs, the execution semantics should be preserved in the value and time domains, and the task implementation should provide modular and reusable code. Previous research on component models for code generation did not consider the execution time of block implementations and the time at which outputs are produced, and did not explore the selection of task generation and scheduling based on output latencies. In this work, we propose formulations and algorithms for synthesizing SBDs into software tasks, while optimizing objectives that include timing (latency), modularity, reusability, and code size.
@inproceedings{dengTaskSynthesisLatencysensitive2014, title = {Task synthesis for latency-sensitive synchronous block diagram}, doi = {10.1109/SIES.2014.6871195}, booktitle = {Proceedings of the 9th {IEEE} {International} {Symposium} on {Industrial} {Embedded} {Systems} ({SIES} 2014)}, author = {Deng, Peng and Zhu, Qi and Di Natale, Marco and Zeng, Haibo}, month = jun, year = {2014}, note = {ISSN: 2150-3117}, keywords = {Optimization, Timing, Software packages, Ports (Computers), Semantics}, pages = {112--121} }
MWSCAS
Model-based synthesis for real-time embedded systems

Qi Zhu

In 2014 IEEE 57th International Midwest Symposium on Circuits and Systems (MWSCAS), Aug 2014

ISSN: 1558-3899

Abs DOI Bib

Model-based design is being increasingly used in the development of real-time embedded control systems due to its capabilities to support early design verification and validation through formal functional models. Similarly as in the case for circuit design, to facilitate the adoption of high level functional models and truly reduce design complexity, it is important to have automated synthesis tools that can generate correct and optimal implementations from those functional models. The development of such synthesis tools has some unique challenges compared to synchronous circuit design - the functional model for real-time embedded systems has more diverse semantics, the implementation platform is more distributed and often asynchronous, and there are often strict timing requirements along with various design objectives such as system performance, safety, security and extensibility. In this paper, we discuss the major challenges in developing model-based synthesis tools for real-time embedded systems, and present an overview of our integrated synthesis flow that addresses task generation, task mapping, and code generation in a holistic fashion. The synthesis process considers a variety of design objectives, and we will highlight the trade-off between timing-related objectives and security.
@inproceedings{zhuModelbasedSynthesisRealtime2014, title = {Model-based synthesis for real-time embedded systems}, doi = {10.1109/MWSCAS.2014.6908428}, booktitle = {2014 {IEEE} 57th {International} {Midwest} {Symposium} on {Circuits} and {Systems} ({MWSCAS})}, author = {Zhu, Qi}, month = aug, year = {2014}, note = {ISSN: 1558-3899}, keywords = {Security, Timing, Real-time systems, Embedded systems, Integrated circuit modeling}, pages = {366--369} }
CODES+ISSS
Metronomy: A function-architecture co-simulation framework for timing verification of cyber-physical systems

Liangpeng Guo, Qi Zhu, Pierluigi Nuzzo, and 3 more authors

In 2014 International Conference on Hardware/Software Codesign and System Synthesis (CODES+ISSS), Oct 2014

Abs DOI Bib

As the design complexity of cyber-physical systems continues to grow, modeling the system at higher abstraction levels with formal models of computation is increasingly appealing since it enables early design verification and analysis. One of the most important aspects in system modeling and analysis is timing. However, it is very challenging to analyze and verify timing at the early design stages, as the design representation is quite abstract and trade-offs have to be made between the performance requirements defined in terms of system functionality and the cost of the feasible architecture that can implement the functionality. In this paper, we present Metronomy, a function-architecture co-simulation framework that integrates functional modeling from Ptolemy and architectural modeling from the MetroII environment via a mapping interface. Metronomy exploits contract theory for timing verification and design space exploration via co-simulation. Two case studies on an electrical power system and a paper-feed sub-system for a high speed printing press demonstrate the effectiveness of our approach.
@inproceedings{guoMetronomyFunctionarchitectureCosimulation2014, title = {Metronomy: {A} function-architecture co-simulation framework for timing verification of cyber-physical systems}, shorttitle = {Metronomy}, doi = {10.1145/2656075.2656093}, booktitle = {2014 {International} {Conference} on {Hardware}/{Software} {Codesign} and {System} {Synthesis} ({CODES}+{ISSS})}, author = {Guo, Liangpeng and Zhu, Qi and Nuzzo, Pierluigi and Passerone, Roberto and Sangiovanni-Vincentelli, Alberto and Lee, Edward A.}, month = oct, year = {2014}, keywords = {Computer architecture, Sensors, Computational modeling, Timing, Co-simulation, Contracts, Cyber-Physical System, Solid modeling, TV}, pages = {1--10} }
ICCAD
Security-aware mapping for TDMA-based real-time distributed systems

Chung-Wei Lin, Qi Zhu, and Alberto Sangiovanni-Vincentelli

In 2014 IEEE/ACM International Conference on Computer-Aided Design (ICCAD), Nov 2014

ISSN: 1558-2434

Abs DOI Bib

Cyber-security has become a critical issue for realtime distributed embedded systems in domains such as automotive, avionics, and industrial automation. However, in many of such systems, tight resource constraints and strict timing requirements make it difficult or even impossible to add security mechanisms after the initial design stages. To produce secure and safe systems with desired performance, security must be considered together with other objectives at the system level and from the beginning of the design. In this paper, we focus on security-aware design for Time Division Multiple Access (TDMA) based real-time distributed systems. The TDMA-based protocol we consider is an abstraction of many time-triggered protocols that are being adopted in various safety-critical systems for their more predictable timing behavior, such as FlexRay, Time-Triggered Protocol, and Time-Triggered Ethernet. To protect against attacks on TDMA-based real-time distributed systems, we apply a message authentication mechanism with time-delayed release of keys, which provides a good balance between security and computational overhead but needs sophisticated network scheduling to ensure that the increased latencies due to delayed key releases will not violate timing requirements. We propose formulations and an algorithm to optimize the task allocation, priority assignment, network scheduling, and key-release interval length during the mapping process, while meeting both security and timing requirements. Experimental results of an automotive case study and a synthetic example show the effectiveness and efficiency of our approach.
@inproceedings{linSecurityawareMappingTDMAbased2014, title = {Security-aware mapping for {TDMA}-based real-time distributed systems}, doi = {10.1109/ICCAD.2014.7001325}, booktitle = {2014 {IEEE}/{ACM} {International} {Conference} on {Computer}-{Aided} {Design} ({ICCAD})}, author = {Lin, Chung-Wei and Zhu, Qi and Sangiovanni-Vincentelli, Alberto}, month = nov, year = {2014}, note = {ISSN: 1558-2434}, keywords = {Security, Resource management, Protocols, Delays, Real-time systems, Receivers}, pages = {24--31} }
ICCAD
Lifetime optimization for real-time embedded systems considering electromigration effects

Taeyoung Kim, Bowen Zheng, Hai-Bao Chen, and 3 more authors

In 2014 IEEE/ACM International Conference on Computer-Aided Design (ICCAD), Nov 2014

ISSN: 1558-2434

Abs DOI Bib

In this article, we propose a new lifetime task optimization technique for real-time embedded processors considering the electromigration-induced reliability. The new approach is based on a recently proposed physics-based electromigration (EM) model for more accurate EM assessment of a power grid network at the chip level. We apply the dynamic voltage and frequency scaling (DVFS) (by selecting the performance states or p-states of the tasks to manage the power) and thus the lifetime of the processor running different tasks over their periods. We consider both single-rate and multi-rate embedded systems with preemption. To model the mean-time-to-failure (MTTF) of a task for a given p-state, response surface modeling is applied. We then frame the reliability optimization problem as the continuous constrained nonlinear optimization problem in which the system EM-induced reliability is maximized subject to the timing constraints, which is further solved by simulated annealing method. Experimental results show that for low utilization systems, significant reliability improvement can be achieved with even smaller power consumption than existing reliability-ignore scheduling method. The proposed method can lead to near Pareto’s front trade-off between the power/energy and the lifetime compared to the existing task scheduling method.
@inproceedings{kimLifetimeOptimizationRealtime2014, title = {Lifetime optimization for real-time embedded systems considering electromigration effects}, doi = {10.1109/ICCAD.2014.7001388}, booktitle = {2014 {IEEE}/{ACM} {International} {Conference} on {Computer}-{Aided} {Design} ({ICCAD})}, author = {Kim, Taeyoung and Zheng, Bowen and Chen, Hai-Bao and Zhu, Qi and Sukharev, Valeriy and Tan, Sheldon X.-D.}, month = nov, year = {2014}, note = {ISSN: 1558-2434}, keywords = {Wires, Reliability, Optimization, Stress, Real-time systems, Program processors, Embedded systems}, pages = {434--439} }
ICCAD
Co-scheduling of HVAC control, EV charging and battery usage for building energy efficiency

Tianshu Wei, Qi Zhu, and Mehdi Maasoumy

In 2014 IEEE/ACM International Conference on Computer-Aided Design (ICCAD), Nov 2014

ISSN: 1558-2434

Abs DOI Bib

Building stock consumes 40% of primary energy consumption in the United States. Among various types of energy loads in buildings, HVAC (heating, ventilation, and air conditioning) and EV (electric vehicle) charging are two of the most important ones and have distinct characteristics. HVAC system accounts for 50% of the building energy consumption and typically operates throughout the day, while EV charging is an emerging major energy load that is hard to predict and may cause spikes in energy demand. To maximize building energy efficiency and grid stability, it is important to address both types of energy loads in a holistic framework. Furthermore, on the supply side, the utilization of multiple energy sources such as grid electricity, solar, wind, and battery storage provides more opportunities for energy efficiency, and should be considered together with the scheduling of energy loads. In this paper, we present a novel model predictive control (MPC) based algorithm to co-schedule HVAC control, EV scheduling and battery usage for reducing the total building energy consumption and the peak energy demand, while maintaining the temperature within the comfort zone for building occupants and meeting the deadlines for EV charging. Experiment results demonstrate the effectiveness of our approach under a variety of demand, supply and environment constraints.
@inproceedings{weiCoschedulingHVACControl2014, title = {Co-scheduling of {HVAC} control, {EV} charging and battery usage for building energy efficiency}, doi = {10.1109/ICCAD.2014.7001351}, booktitle = {2014 {IEEE}/{ACM} {International} {Conference} on {Computer}-{Aided} {Design} ({ICCAD})}, author = {Wei, Tianshu and Zhu, Qi and Maasoumy, Mehdi}, month = nov, year = {2014}, note = {ISSN: 1558-2434}, keywords = {Prediction algorithms, Power grids, Buildings, Mathematical model, Energy consumption, Equations, Batteries}, pages = {191--196} }

2013

Optimization of task allocation and priority assignment in hard real-time distributed systems

Qi Zhu, Haibo Zeng, Wei Zheng, and 2 more authors

ACM Transactions on Embedded Computing Systems, Jan 2013

Abs DOI Bib HTML

The complexity and physical distribution of modern active safety, chassis, and powertrain automotive applications requires the use of distributed architectures. Complex functions designed as networks of function blocks exchanging signal information are deployed onto the physical HW and implemented in a SW architecture consisting of a set of tasks and messages. The typical configuration features priority-based scheduling of tasks and messages and imposes end-to-end deadlines. In this work, we present and compare formulations and procedures for the optimization of the task allocation, the signal to message mapping, and the assignment of priorities to tasks and messages in order to meet end-to-end deadline constraints and minimize latencies. Our formulations leverage worst-case response time analysis within a mixed integer linear optimization framework and are compared for performance against a simulated annealing implementation. The methods are applied for evaluation to an automotive case study of complexity comparable to industrial design problems.
@article{zhuOptimizationTaskAllocation2013, title = {Optimization of task allocation and priority assignment in hard real-time distributed systems}, volume = {11}, issn = {1539-9087}, url = {https://doi.org/10.1145/2362336.2362352}, doi = {10.1145/2362336.2362352}, number = {4}, urldate = {2022-05-06}, journal = {ACM Transactions on Embedded Computing Systems}, author = {Zhu, Qi and Zeng, Haibo and Zheng, Wei and Natale, Marco DI and Sangiovanni-Vincentelli, Alberto}, month = jan, year = {2013}, keywords = {automotive systems, Design optimization, architectures, optimization, real-time systems, schedulability}, pages = {85:1--85:30} }
metro II: A design environment for cyber-physical systems

Abhijit Davare, Douglas Densmore, Liangpeng Guo, and 4 more authors

ACM Transactions on Embedded Computing Systems, Mar 2013

Abs DOI Bib HTML

Cyber-Physical Systems are integrations of computation and physical processes and as such, will be increasingly relevant to industry and people. The complexity of designing CPS resides in their heterogeneity. Heterogeneity manifest itself in modeling their functionality as well as in the implementation platforms that include a multiplicity of components such as microprocessors, signal processors, peripherals, memories, sensors and actuators often integrated on a single chip or on a small package such as a multi-chip module. We need a methodology, tools and environments where heterogeneity can be dealt with at all levels of abstraction and where different tools can be integrated. We present here Platform-Based Design as the CPS methodology of choice and metroII, a design environment that supports it. We present the metamodeling approach followed in metroII, how to couple the functionality and implementation platforms of CPS, and the simulation technology that supports the analysis of CPS and of their implementation. We also present examples of use and the integration of metroII with another popular design environment developed at Verimag, BIP.
@article{davareMetroIIDesign2013, title = {metro {II}: {A} design environment for cyber-physical systems}, volume = {12}, issn = {1539-9087}, shorttitle = {{\textless}span class="smallcaps {smallerCapital}"{\textgreater}metro{\textless}/span{\textgreater}{II}}, url = {https://doi.org/10.1145/2435227.2435245}, doi = {10.1145/2435227.2435245}, number = {1s}, urldate = {2022-05-06}, journal = {ACM Transactions on Embedded Computing Systems}, author = {Davare, Abhijit and Densmore, Douglas and Guo, Liangpeng and Passerone, Roberto and Sangiovanni-Vincentelli, Alberto L. and Simalatsar, Alena and Zhu, Qi}, month = mar, year = {2013}, keywords = {Platform-Based Design, Cyber-Physical Systems, Heterogeneous Embedded Systems, Modeling, Multiprocessor, System-on-Chip}, pages = {49:1--49:31} }
DATE
Robust and extensible task implementations of synchronous finite state machines

Qi Zhu, Peng Deng, Marco Di Natale, and 1 more author

In 2013 Design, Automation Test in Europe Conference Exhibition (DATE), Mar 2013

ISSN: 1530-1591

Abs DOI Bib

Model-based design using synchronous reactive (SR) models is widespread for the development of embedded control software. SR models ease verification and validation, and enable the automatic generation of implementations. In SR models, synchronous finite state machines (FSMs) are commonly used to capture changes of the system state under trigger events. The implementation of a synchronous FSM may be improved by using multiple software tasks instead of the traditional single-task solution. In this work, we propose methods to quantitatively analyze task implementations with respect to a breakdown factor that measures the timing robustness, and an action extensibility metric that measures the capability to accommodate upgrades. We propose an algorithm to generate a correct and efficient task implementation of synchronous FSMs for these two metrics, while guaranteeing the schedulability constraints.
@inproceedings{zhuRobustExtensibleTask2013, title = {Robust and extensible task implementations of synchronous finite state machines}, doi = {10.7873/DATE.2013.272}, booktitle = {2013 {Design}, {Automation} {Test} in {Europe} {Conference} {Exhibition} ({DATE})}, author = {Zhu, Qi and Deng, Peng and Di Natale, Marco and Zeng, Haibo}, month = mar, year = {2013}, note = {ISSN: 1530-1591}, keywords = {Algorithm design and analysis, Computational modeling, Electric breakdown, Measurement, Partitioning algorithms, Software packages}, pages = {1319--1324} }
ICCPS
Co-design of control algorithm and embedded platform for building HVAC systems

Mehdi Maasoumy, Qi Zhu, Cheng Li, and 2 more authors

In 2013 ACM/IEEE International Conference on Cyber-Physical Systems (ICCPS), Apr 2013

Abs DOI Bib

The design of heating, ventilation and air conditioning (HVAC) systems is crucial for reducing energy consumption in buildings. As complex cyber-physical systems, HVAC systems involve three closely-related subsystems — the control algorithm, the physical building and environment and the embedded implementation platform. In the traditional top-down approach, the control algorithm and the embedded platform are in general designed separately leading to suboptimal systems. We propose a co-design approach that analyzes the interaction between the control algorithm and the embedded platform through a set of interface variables (in this paper we address in particular sensing accuracy). We present six control algorithms that take into account the sensing error, and model the relation of control performance and cost versus sensing error. We also capture the relation of embedded platform cost versus sensing error by analysis of the collected data from a test bed. Based on these models, we explore the co-design of the control algorithm and the temperature sensing subsystem of the embedded platform to optimize with respect to energy cost and monetary cost while satisfying the constraints for user comfort level.
@inproceedings{maasoumyCodesignControlAlgorithm2013, title = {Co-design of control algorithm and embedded platform for building {HVAC} systems}, doi = {10.1145/2502524.2502533}, booktitle = {2013 {ACM}/{IEEE} {International} {Conference} on {Cyber}-{Physical} {Systems} ({ICCPS})}, author = {Maasoumy, Mehdi and Zhu, Qi and Li, Cheng and Meggers, Forrest and Sangiovanni-Vincentelli, Alberto}, month = apr, year = {2013}, keywords = {Algorithm design and analysis, Buildings, Mathematical model, Accuracy, building energy efficiency, co-design, Temperature measurement, Temperature sensors}, pages = {61--70} }
ICCAD
Security-aware mapping for CAN-based real-time distributed automotive systems

Chung-Wei Lin, Qi Zhu, Calvin Phung, and 1 more author

In 2013 IEEE/ACM International Conference on Computer-Aided Design (ICCAD), Nov 2013

ISSN: 1558-2434

Abs DOI Bib

Cyber-security is a rising issue for automotive electronic systems, and it is critical to system safety and dependability. Current in-vehicles architectures, such as those based on the Controller Area Network (CAN), do not provide direct support for secure communications. When retrofitting these architectures with security mechanisms, a major challenge is to ensure that system safety will not be hindered, given the limited computation and communication resources. We apply Message Authentication Codes (MACs) to protect against masquerade and replay attacks on CAN networks, and propose an optimal Mixed Integer Linear Programming (MILP) formulation for solving the mapping problem from a functional model to the CAN-based platform while meeting both the security and the safety requirements. We also develop an efficient heuristic for the mapping problem under security and safety constraints. To the best of our knowledge, this is the first work to address security and safety in an integrated formulation in the design automation of automotive electronic systems. Experimental results of an industrial case study show the effectiveness of our approach.
@inproceedings{linSecurityawareMappingCANbased2013, title = {Security-aware mapping for {CAN}-based real-time distributed automotive systems}, doi = {10.1109/ICCAD.2013.6691106}, booktitle = {2013 {IEEE}/{ACM} {International} {Conference} on {Computer}-{Aided} {Design} ({ICCAD})}, author = {Lin, Chung-Wei and Zhu, Qi and Phung, Calvin and Sangiovanni-Vincentelli, Alberto}, month = nov, year = {2013}, note = {ISSN: 1558-2434}, keywords = {Security, Resource management, Safety, Mathematical model, Equations, Receivers, Time factors}, pages = {115--121} }

2012

Development of Building Automation and Control Systems

Yang Yang, Qi Zhu, Mehdi Maasoumy, and 1 more author

IEEE Design Test of Computers, Aug 2012

Conference Name: IEEE Design Test of Computers

Abs DOI Bib

In this paper, we proposed a design flow for BAC systems that enables integrating heterogeneous input models, conducts automatic design space exploration, and performs software synthesis on distributed platforms while guaranteeing correctness and reducing communication load. We believe these capabilities can enable the building designers to better adopt model-based design methodologies, and facilitate them to improve design productivity, optimize system performance, and reduce cost.

@article{yangDevelopmentBuildingAutomation2012,
  title = {Development of {Building} {Automation} and {Control} {Systems}},
  volume = {29},
  issn = {1558-1918},
  doi = {10.1109/MDT.2012.2201130},
  number = {4},
  journal = {IEEE Design Test of Computers},
  author = {Yang, Yang and Zhu, Qi and Maasoumy, Mehdi and Sangiovanni-Vincentelli, Alberto},
  month = aug,
  year = {2012},
  note = {Conference Name: IEEE Design Test of Computers},
  keywords = {Computer architecture, Control systems, Mathematical model, Computational modeling, Software engineering, Building Automation and Control, design space exploration, Green buildings, intermediate format, Program processors, software synthesis},
  pages = {45--55}
}

ETFA
Optimizing stack memory requirements for real-time embedded applications

Haibo Zeng, Marco Di Natale, and Qi Zhu

In Proceedings of 2012 IEEE 17th International Conference on Emerging Technologies Factory Automation (ETFA 2012), Sep 2012

ISSN: 1946-0759

Abs DOI Bib

In the development of some real-time embedded applications, especially systems-on-chip, an efficient use of RAM memory is as important as the effective scheduling of the computation resources. The design problem is to find a schedulable solution that fits within the memory budget. In a real-time concurrent system, preemption plays an important role in the exploration of these tradeoffs. Several schemes, including preemption thresholds and non-preemption groups, have been developed to improve schedulability and saving stack memory space by selectively disabling preemption. However, the design synthesis problem for such systems and protocols is still an open problem. We target at the efficient assignment of the scheduling parameters for systems scheduled according to these policies in several cases of practical interest, including those that are compliant with automotive standards.
@inproceedings{zengOptimizingStackMemory2012, title = {Optimizing stack memory requirements for real-time embedded applications}, doi = {10.1109/ETFA.2012.6489571}, booktitle = {Proceedings of 2012 {IEEE} 17th {International} {Conference} on {Emerging} {Technologies} {Factory} {Automation} ({ETFA} 2012)}, author = {Zeng, Haibo and Di Natale, Marco and Zhu, Qi}, month = sep, year = {2012}, note = {ISSN: 1946-0759}, pages = {1--8} }

2010

Optimizing the Software Architecture for Extensibility in Hard Real-Time Distributed Systems

Qi Zhu, Yang Yang, Marco Natale, and 2 more authors

IEEE Transactions on Industrial Informatics, Nov 2010

Conference Name: IEEE Transactions on Industrial Informatics

Abs DOI Bib

We consider a set of control tasks that must be executed on distributed platforms so that end-to-end latencies are within deadlines. We investigate how to allocate tasks to nodes, pack signals to messages, allocate messages to buses, and assign priorities to tasks and messages, so that the design is extensible and robust with respect to changes in task requirements. We adopt a notion of extensibility metric that measures how much the execution times of tasks can be increased without violating end-to-end deadlines. We optimize the task and message design with respect to this metric by adopting a mathematical programming front-end followed by postprocessing heuristics. The proposed algorithm as applied to industrial strength test cases shows its effectiveness in optimizing extensibility and a marked improvement in running time with respect to an approach based on randomized optimization.
@article{zhuOptimizingSoftwareArchitecture2010, title = {Optimizing the {Software} {Architecture} for {Extensibility} in {Hard} {Real}-{Time} {Distributed} {Systems}}, volume = {6}, issn = {1941-0050}, doi = {10.1109/TII.2010.2053938}, number = {4}, journal = {IEEE Transactions on Industrial Informatics}, author = {Zhu, Qi and Yang, Yang and Natale, Marco and Scholte, Eelco and Sangiovanni-Vincentelli, Alberto}, month = nov, year = {2010}, note = {Conference Name: IEEE Transactions on Industrial Informatics}, keywords = {Real time systems, Robustness, Control systems, Sensor systems, Design optimization, Automatic control, Control system synthesis, Design space exploration, distributed system, extensibility, platform-based design, Production, real-time, Software architecture, Supply chains}, pages = {621--636} }
A Design Flow for Building Automation and Control Systems

Yang Yang, Alessandro Pinto, Alberto Sangiovanni-Vincentelli, and 1 more author

In 2010 31st IEEE Real-Time Systems Symposium, Nov 2010

ISSN: 1052-8725

Abs DOI Bib

We propose a system-level design flow for building automation and control (BAC) systems. The input to the design flow is a high level description of the control algorithms given in a model-based environment such as Simulink. The input specification is translated into an intermediate format, and then automatically refined into a distributed implementation. Refinement includes optimal mapping of the functional specification on a set of computation and communication resources, and software synthesis, which generates code for each component in the mapped design while guaranteeing semantic equivalence with the original specification. Experiments with a temperature control system are presented to illustrate the flow.
@inproceedings{yangDesignFlowBuilding2010, title = {A {Design} {Flow} for {Building} {Automation} and {Control} {Systems}}, doi = {10.1109/RTSS.2010.26}, booktitle = {2010 31st {IEEE} {Real}-{Time} {Systems} {Symposium}}, author = {Yang, Yang and Pinto, Alessandro and Sangiovanni-Vincentelli, Alberto and Zhu, Qi}, month = nov, year = {2010}, note = {ISSN: 1052-8725}, keywords = {Computer architecture, Algorithm design and analysis, Buildings, Control systems, Computational modeling, Libraries, Atmospheric modeling}, pages = {105--116} }

2009

Optimizing Extensibility in Hard Real-Time Distributed Systems

Qi Zhu, Yang Yang, Eelco Scholte, and 2 more authors

In 2009 15th IEEE Real-Time and Embedded Technology and Applications Symposium, Apr 2009

ISSN: 1545-3421

Abs DOI Bib

We consider a set of control tasks that must be executed on distributed platforms so that end-to-end latencies are within deadlines. We investigate how to allocate tasks to nodes, pack signals to messages, allocate messages to buses, and assign priorities to tasks and messages, so that the design is robust with respect to changes in task requirements. The notion of extensibility is used to measure robustness. The extensibility metric measures how much the execution times of tasks can be increased without violating end-to-end deadlines. We optimize this metric by adopting a mathematical programming front-end followed by post-processing heuristics. The proposed algorithm as applied to industrial strength test cases shows its effectiveness in optimizing extensibility and a marked improvement in running time with respect to an approach based on randomized optimization.
@inproceedings{zhuOptimizingExtensibilityHard2009, title = {Optimizing {Extensibility} in {Hard} {Real}-{Time} {Distributed} {Systems}}, doi = {10.1109/RTAS.2009.37}, booktitle = {2009 15th {IEEE} {Real}-{Time} and {Embedded} {Technology} and {Applications} {Symposium}}, author = {Zhu, Qi and Yang, Yang and Scholte, Eelco and Natale, Marco Di and Sangiovanni-Vincentelli, Alberto}, month = apr, year = {2009}, note = {ISSN: 1545-3421}, keywords = {Computer architecture, Real time systems, Robustness, Control systems, Optimization, Sensor systems, Actuators, Delay, Constraint optimization, Design optimization, Extensibility, Platform-Based Design, Real-time Distributed Systems, Signal design}, pages = {275--284} }

2007

Period Optimization for Hard Real-time Distributed Automotive Systems

Abhijit Davare, Qi Zhu, Marco Di Natale, and 3 more authors

In 2007 44th ACM/IEEE Design Automation Conference, Jun 2007

ISSN: 0738-100X

Abs Bib

The complexity and physical distribution of modern active-safety automotive applications requires the use of distributed architectures. These architectures consist of multiple electronic control units (ECUs) connected with standardized buses. The most common configuration features periodic activation of tasks and messages coupled with run-time priority-based scheduling. The correct deployment of applications on such architectures requires end-to- end latency deadlines to be met. This is challenging since deadlines must be enforced across a set of ECUs and buses, each of which supports multiple functionality. The need for accommodating legacy tasks and messages further complicates the scenario. In this work, we automatically assign task and message periods for distributed automotive systems. This is accomplished by leveraging schedulability analysis within a convex optimization framework to simultaneously assign periods and satisfy end-to-end latency constraints. Our approach is applied to an industrial case study as well as an example taken from the literature and is shown to be both effective and efficient.
@inproceedings{davarePeriodOptimizationHard2007, title = {Period {Optimization} for {Hard} {Real}-time {Distributed} {Automotive} {Systems}}, booktitle = {2007 44th {ACM}/{IEEE} {Design} {Automation} {Conference}}, author = {Davare, Abhijit and Zhu, Qi and Di Natale, Marco and Pinello, Claudio and Kanajan, Sri and Sangiovanni-Vincentelli, Alberto}, month = jun, year = {2007}, note = {ISSN: 0738-100X}, keywords = {Aerospace electronics, Real time systems, Job shop scheduling, Automotive engineering, Clocks, Timing, Runtime, Algorithms, activation period, Actuators, automotive systems, Delay, end-to-end latency, Performance, Permission}, pages = {278--283} }
RTSS
Definition of Task Allocation and Priority Assignment in Hard Real-Time Distributed Systems

Wei Zheng, Qi Zhu, Marco Di Natale, and 1 more author

In 28th IEEE International Real-Time Systems Symposium (RTSS 2007), Dec 2007

ISSN: 1052-8725

Abs DOI Bib

The complexity and physical distribution of modern active safety, chassis and powertrain automotive applications requires the use of distributed architectures. Complex functions designed as networks of function blocks exchanging signal information are deployed onto the physical HW and implemented in a SW architecture consisting of a set of tasks and messages. The typical configuration features priority-based scheduling of tasks and messages and imposes end- to-end deadlines. In this work, we optimize the task placement and the signal to message mapping and we automate the assignment of priorities to tasks and messages in order to meet end-to-end deadline constraints and minimize latencies. This is accomplished by leveraging worst case response time analysis within a mixed integer linear optimization framework. Our approach is applied to an automotive case study to prove its feasibility.
@inproceedings{zhengDefinitionTaskAllocation2007, title = {Definition of {Task} {Allocation} and {Priority} {Assignment} in {Hard} {Real}-{Time} {Distributed} {Systems}}, doi = {10.1109/RTSS.2007.40}, booktitle = {28th {IEEE} {International} {Real}-{Time} {Systems} {Symposium} ({RTSS} 2007)}, author = {Zheng, Wei and Zhu, Qi and Natale, Marco Di and Vincentelli, Alberto Sangiovanni}, month = dec, year = {2007}, note = {ISSN: 1052-8725}, keywords = {Real time systems, Safety, Automotive engineering, Timing, Delay, Automotive applications, Communication system control, Mechanical power transmission, Processor scheduling, Signal mapping}, pages = {161--170} }
A Next-Generation Design Framework for Platform-Based Design

Abhijit Davare, Douglas Densmore, Trevor Meyerowitz, and 5 more authors

In 2007 Design and Verification Conference, Feb 2007

Abs Bib

The platform-based design methodology [1] is based on the usage of formal modeling techniques, clearly deﬁned abstraction levels and the separation of concerns to enable an effective design process. The METROPOLIS framework embodies the platform-based design methodology and has been applied to a number of case studies across multiple domains. Based on these experiences, we have identiﬁed three key features that need to be enhanced: heterogeneous IP import, orthogonalization of performance from behavior, and design space exploration. The next generation METRO II framework incorporates these advanced features. The main concepts underlying METRO II are described in this paper and illustrated with a small example.
@inproceedings{davareNextGenerationDesignFramework2007, address = {San Jose, CA}, title = {A {Next}-{Generation} {Design} {Framework} for {Platform}-{Based} {Design}}, language = {en}, booktitle = {2007 {Design} and {Verification} {Conference}}, author = {Davare, Abhijit and Densmore, Douglas and Meyerowitz, Trevor and Pinto, Alessandro and Sangiovanni-Vincentelli, Alberto and Yang, Guang and Zeng, Haibo and Zhu, Qi}, month = feb, year = {2007} }

2006

SAT sweeping with local observability don’t-cares

Qi Zhu, N. Kitchen, A. Kuehlmann, and 1 more author

In 2006 43rd ACM/IEEE Design Automation Conference, Jul 2006

ISSN: 0738-100X

Abs DOI Bib

SAT sweeping is a method for simplifying an AND/inverter graph (AIG) by systematically merging graph vertices from the inputs towards the outputs using a combination of structural hashing, simulation, and SAT queries. Due to its robustness and efficiency, SAT sweeping provides a solid algorithm for Boolean reasoning in functional verification and logic synthesis. In previous work, SAT sweeping merges two vertices only if they are functionally equivalent. In this paper we present a significant extension of the SAT-sweeping algorithm that exploits local observability don’t-cares (ODCs) to increase the number of vertices merged. We use a novel technique to bound the use of ODCs and thus the computational effort to find them, while still finding a large fraction of them. Our reported results based on a set of industrial benchmark circuits demonstrate that ODC-based SAT sweeping results in significantly more graph simplification with great benefit for Boolean reasoning with a moderate increase in computational effort.
@inproceedings{zhuSATSweepingLocal2006, title = {{SAT} sweeping with local observability don't-cares}, doi = {10.1109/DAC.2006.229206}, booktitle = {2006 43rd {ACM}/{IEEE} {Design} {Automation} {Conference}}, author = {Zhu, Qi and Kitchen, N. and Kuehlmann, A. and Sangiovanni-Vincentelli, A.}, month = jul, year = {2006}, note = {ISSN: 0738-100X}, keywords = {Robustness, Merging, Observability, Computational modeling, Algorithms, And/inverter graphs, Boolean functions, Circuit synthesis, Data structures, Inverters, Logic design, observability, SAT sweeping, Solids, Verification}, pages = {229--234} }

2005

ASAP
Via-aware global routing for good VLSI manufacturability and high yield

Yang Yang, Tong Jing, Xianlong Hong, and 4 more authors

In 2005 IEEE International Conference on Application-Specific Systems, Architecture Processors (ASAP’05), Jul 2005

ISSN: 1063-6862

Abs DOI Bib

CAD tools have become more and more important for integrated circuit (IC) design since a complicated system can be designed into a single chip, called system-on-a-chip (SOC), in which physical design tool is an essential and critical part. We try to consider the via minimization problem as early as possible in physical design. We propose a routing method focusing on minimizing vias while considering mutability and wire-length constraint. That is, in the global routing phase, we minimize the number of bends, which is closely related to the number of vias. Previous work only dealt with very small nets, but our algorithm is general for the nets with any size. Experimental results show that our algorithm can greatly reduce the count of bends for various sizes of nets while meeting the constraints of congestion and wire-length.
@inproceedings{yangViaawareGlobalRouting2005, title = {Via-aware global routing for good {VLSI} manufacturability and high yield}, doi = {10.1109/ASAP.2005.67}, booktitle = {2005 {IEEE} {International} {Conference} on {Application}-{Specific} {Systems}, {Architecture} {Processors} ({ASAP}'05)}, author = {Yang, Yang and Jing, Tong and Hong, Xianlong and Hu, Yu and Zhu, Qi and Hu, Xiaodong and Yan, Guiying}, month = jul, year = {2005}, note = {ISSN: 1063-6862}, keywords = {Computer science, Design automation, Circuit optimization, Computer aided manufacturing, Content addressable storage, Integrated circuit technology, Minimization, Routing, System-on-a-chip, Very large scale integration}, pages = {198--203} }
Spanning graph-based nonrectilinear steiner tree algorithms

Qi Zhu, Hai Zhou, Tong Jing, and 2 more authors

IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems, Jul 2005

Conference Name: IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems

Abs DOI Bib

With advances in fabrication technology of very/ultra large scale integrated circuit (VLSI/ULSI), we must face many new challenges. One of them is the interconnect effects, which may cause longer delay and heavier crosstalk. To solve this problem, many interconnect performance optimization algorithms have been proposed. However, when these algorithms are designed based on rectilinear interconnect architecture, the optimization capability is limited. Therefore, nonrectilinear interconnect architectures become a field of active research in which the octilinear interconnect architecture is the most promising one since it extends from the rectilinear case and greatly shortens the wire length. Meanwhile, an interconnect with less length is helpful to reduce wire capacitance, congestion, and routing area. In an interconnect architecture, the Steiner minimal tree (SMT) construction is one of the key problems. In this paper, we give two practical octilinear Steiner minimal tree (OSMT) construction algorithms based on octilinear spanning graphs (OSGs). The one with edge substitution (OST-E) has a worst-case running time of O(nlogn) and a similar performance as the recent work using batched greedy. The other one with triangle contraction (OST-T) has a small increase in the constant factor of running time and a better performance. These two are the fastest algorithms for octilinear Steiner tree construction so far. Experiments on both industrial and random test cases are conducted to compare with other programs. We also proposed the extension of our algorithms to any /spl lambda/ geometry.
@article{zhuSpanningGraphbasedNonrectilinear2005, title = {Spanning graph-based nonrectilinear steiner tree algorithms}, volume = {24}, issn = {1937-4151}, doi = {10.1109/TCAD.2005.850862}, number = {7}, journal = {IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems}, author = {Zhu, Qi and Zhou, Hai and Jing, Tong and Hong, X.-L. and Yang, Yang}, month = jul, year = {2005}, note = {Conference Name: IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems}, keywords = {routing, Steiner trees, Tree graphs, Wire, Integrated circuit technology, Very large scale integration, Crosstalk, Deep submicron (DSM), Delay effects, Fabrication, Integrated circuit interconnections, interconnect, physical design, Steiner tree, Ultra large scale integration, very large scale integration (VLSI)}, pages = {1066--1075} }
JPEG encoding on the Intel MXP5800: a platform-based design case study

A. Davare, Qi Zhu, J. Moondanos, and 1 more author

In 3rd Workshop on Embedded Systems for Real-Time Multimedia, 2005., Sep 2005

Abs DOI Bib

Multimedia systems are becoming increasingly complex and concurrent. The platform-based design (PBD) methodology (Keutzer et al., 2000) tackles these issues by recommending the use of formal models, carefully defined abstraction layers and the separation of concerns. Models of computation (Lee and Sangiovanni-Vincentelli, 1998) (MoCs) can be used within this methodology to enable specialized synthesis and verification techniques. In this paper, these concepts are leveraged in an industrial case study: the JPEG encoder application deployed on the Intel MXP5800 imaging processor. The modeling is carried out in the Metropolis (Balarin et al., 2003) design framework. We show that the system-level model using our chosen model of computation allows performance estimation within 5% of the actual implementation. Moreover, the chosen MoC is amenable to automation, which enables future synthesis techniques.
@inproceedings{davareJPEGEncodingIntel2005, title = {{JPEG} encoding on the {Intel} {MXP5800}: a platform-based design case study}, shorttitle = {{JPEG} encoding on the {Intel} {MXP5800}}, doi = {10.1109/ESTMED.2005.1518081}, booktitle = {3rd {Workshop} on {Embedded} {Systems} for {Real}-{Time} {Multimedia}, 2005.}, author = {Davare, A. and Zhu, Qi and Moondanos, J. and Sangiovanni-Vincentelli, A.}, month = sep, year = {2005}, keywords = {Streaming media, Computational modeling, Computer aided software engineering, Design methodology, Discrete cosine transforms, Encoding, Multimedia systems, Quantization, Space exploration, Transform coding}, pages = {89--94} }

2004

ASP-DAC
Efficient octilinear steiner tree construction based on spanning graphs

Qi Zhu, Hai Zhou, Tong Jing, and 2 more authors

In ASP-DAC 2004: Asia and South Pacific Design Automation Conference 2004 (IEEE Cat. No.04EX753), Jan 2004

Abs DOI Bib

0ctilinear interconnect is a promising technique to shorten wire lengths. We present two practical heuristic octilinear Steiner tree (OSMT) algorithms in the paper. They are both based on octilinear spanning graphs. The one by edge substitution (OST-E) has a worst case running time of O(nlogn) and similar performance as the batched greedy algorithm[9]. The other one by triangle contraction (OST-T) has a small increase in running time and better performance. Experiments on both industry and random test cases are conducted.
@inproceedings{zhuEfficientOctilinearSteiner2004, title = {Efficient octilinear steiner tree construction based on spanning graphs}, doi = {10.1109/ASPDAC.2004.1337680}, booktitle = {{ASP}-{DAC} 2004: {Asia} and {South} {Pacific} {Design} {Automation} {Conference} 2004 ({IEEE} {Cat}. {No}.{04EX753})}, author = {Zhu, Qi and Zhou, Hai and Jing, Tong and Hong, Xianlong and Yang, Yang}, month = jan, year = {2004}, keywords = {Computer architecture, Heuristic algorithms, Construction industry, Greedy algorithms, Postal services, Steiner trees, Surface-mount technology, Testing, Tree graphs, Wire}, pages = {687--690} }

2003

ICT
Buffer replacement algorithm for merge-based multicast video-on-demand system

Qi Zhu, Ling Shao, Rong Yan, and 2 more authors

In 10th International Conference on Telecommunications, 2003. ICT 2003., Feb 2003

Abs DOI Bib

The traditional buffer replacement algorithms do not perform well in the multicast video-on-demand (VoD) systems. In this paper, we propose a new algorithm, named urgent overlap frequency (UOF), for merge-based multicast VoD servers. It makes good use of the merge property of the multicast channels. Unlike traditional ones that consider time urgent degree or access frequency, three factors, i.e. time urgent degree, access overlap and access frequency, are considered in UOF algorithm. The experimental results show that UOF can increase the hit ratio 90% over LRU/MRU and 20% over BASIC algorithm in multicast circumstance. Additionally, the hit ratio of the UOF algorithm is only 4% lower than the theoretic optimal result, so it is very suitable for merge-based multicast VoD system.
@inproceedings{zhuBufferReplacementAlgorithm2003, title = {Buffer replacement algorithm for merge-based multicast video-on-demand system}, volume = {2}, doi = {10.1109/ICTEL.2003.1191647}, booktitle = {10th {International} {Conference} on {Telecommunications}, 2003. {ICT} 2003.}, author = {Zhu, Qi and Shao, Ling and Yan, Rong and Zhang, Jian and Xie, Dong}, month = feb, year = {2003}, keywords = {Computer science, Merging, Bandwidth, Frequency, Multicast algorithms, Network servers, Partial response channels, Unicast}, pages = {1448--1451 vol.2} }

Can We Trust Embodied Agents? Exploring Backdoor Attacks against Embodied LLM-Based Decision-Making Systems

Ruochen Jiao, Shaoyuan Xie, Justin Yue, and 5 more authors

In The Thirteenth International Conference on Learning Representations

Bib

@inproceedings{jiaocan,
  title = {Can We Trust Embodied Agents? Exploring Backdoor Attacks against Embodied LLM-Based Decision-Making Systems},
  author = {Jiao, Ruochen and Xie, Shaoyuan and Yue, Justin and SATO, TAKAMI and Wang, Lixu and Wang, Yixuan and Chen, Qi Alfred and Zhu, Qi},
  booktitle = {The Thirteenth International Conference on Learning Representations}
}

Directly Forecasting Belief for Reinforcement Learning with Delays

Qingyuan Wu^*, Yuhui Wang^*, Simon Sinong Zhan^*, and 6 more authors

In Forty-second International Conference on Machine Learning

Bib

@inproceedings{wudirectly,
  title = {Directly Forecasting Belief for Reinforcement Learning with Delays},
  author = {Wu, Qingyuan and Wang, Yuhui and Zhan, Simon Sinong and Wang, Yixuan and Lin, Chung-Wei and Lv, Chen and Zhu, Qi and Schmidhuber, J{\"u}rgen and Huang, Chao},
  booktitle = {Forty-second International Conference on Machine Learning}
}

On Large Language Model Continual Unlearning

Chongyang Gao, Lixu Wang, Kaize Ding, and 3 more authors

In The Thirteenth International Conference on Learning Representations

Bib

@inproceedings{gaolarge,
  title = {On Large Language Model Continual Unlearning},
  author = {Gao, Chongyang and Wang, Lixu and Ding, Kaize and Weng, Chenkai and Wang, Xiao and Zhu, Qi},
  booktitle = {The Thirteenth International Conference on Learning Representations}
}