C3PO: Connected Cars & Corridors for Pervasive Sensing and COntrol of Vehicular Flows

VGrid

This research studies the behavior and performance of a transportation corridor made of a freeway and arterial streets under a Connected Vehicles (CV) environment, where vehicles equipped with wireless communication and sensing devices collect, process, and share traffic information among themselves along with roadside sensors. First, methods of fusing/combining real-time traffic data from both vehicles and roadside sensors will be explored to automate the detection of incidents (e.g., accidents) and estimation of the multi-modal traffic demand at intersections or origin-to-destination trip information. Second, the research studies how congestion, particularly traffic jams, emerge and spread, and how cooperative driving technology (CACC), vehicle speed control, and platooning can increase the efficiency of vehicle streams. Our envisioned system will integrate WAVE communications, vehicle positioning, obstacle detection, and in-vehicle processors to form opportunistically high-performance vehicle streams on-demand, particularly at merging, lane-dropping locations and on special lanes in a distributed manner. Finally, the research makes use of the results obtained from the first two tasks to explore algorithms that will enable the adaptive, coordinated control of freeway ramp meters and traffic lights, and the re-routing of traffic in response to traffic incidents.

As part of this project, we have developed Vehicular Network Open Simulator (VENTOS), an integrated C++ simulator that consists of many different modules, including enhancement to SUMO and OMNET.

We hope the results of our research will lead to new ways to monitor and control vehicular traffic, which will enable applications for reducing traffic congestion and fuel consumption. We will build on some of the results and tools developed in our previous collaborative project, VMesh/VGrid, where our team leverages vehicular ad hoc networks (VANET) to perform distributed data sensing, relaying, and computing.

In more recent years, we studied the impact of deep reinforcement learning (DRL)-based traffic signal controller (TSC) on air quality using real traffic demands on city-level road networks. The team also studied the vulnerabilities of these DRL-TSC algorithms in the presence of black-box and white-box adversarial attacks. Our results show that the performance of DRL learning agent decreases in both settings, resulting in higher levels of traffic congestion. We then proposed an ensemble model to perform sequential anomaly detection of the adversarial attacks. Our model minimizes detection delay, achieves lower false alarm rates due to cumulative anomaly inspection.

People

Faculty

Graduate Students

  • Ammar Haydari, Electrical & Computer Engineering (PhD)
  • Jerry Chia-Cheng Yen, Computer Science (PhD)

Alumni

  • Kartik Pandit, Computer Science (PhD, 2013)
  • Hui Deng, Civil & Environmental Engineering (PhD)
  • Mani Amoozadeh, Electrical & Computer Engineering (PhD, 2018)
  • Huajun Chai, Civil & Environmental Engineering (PhD, 2019)
  • Zhongyi Lin, Electrical & Computer Engineering (MS, 2017)
  • Bryan Ching, Electrical & Computer Engineering (MS, 2019)
  • Hasith Rajakarunanay, Electrical & Computer Engineering (B.S., 2018)
  • Zhening Zhang, Electrical & COmputer Engineering (B.S., 2018)

Publications

  • A. Haydari, M. Zhang, and C-N. Chuah, "Adversarial Attacks and Defense in Deep Reinforcement Learning (Deep-RL) Based Traffic Light Controller," IEEE Open Journal of Intelligent Transportation Systems, October 2021. [DOI: 10.1109/OJITS.2021.3118972]
  • A. Haydari, H. M. Zhang, C-N. Chuah, and D. Ghosal, "Impact of Deep RL-based Traffic Signal Control on Air Quality, IEEE Vehicular Technology Conference (VTC 2021)-Spring, April 2021.
  • C-C. Yen, D. Ghosal, M. Zhang, C-N. Chuah, "Security Vulnerabilities and Protection Mechanisms for Backpressure-based Traffic Signal Control," IEEE Transactions on Intelligent Transportation Systems, February 2021. [DOI: 10.1109/TITS.2021.3056658]
  • C-C. Yen, D. Ghosal, H. Michael Zhang, C-N. Chuah, “A Deep On-policy Learning Agent for Traffic Signal Control of Multiple Intersections,” IEEE 23rd International Conference on Intelligent Transportation Systems (ITSC), September 2020.
  • B. Ching, M. Amoozadeh, C-N. Chuah, H. M. Zhang, and D. Ghosal, “Enabling Performance and Security Simulation Studies of Intelligent Traffic Signal Light Control with VENTOS-HIL,” Elsevier Vehicular Communications, vol. 24, August 2020.
  • A. Khodadadi, C-N. Chuah, and T. S. Woo, “Mining Vehicle Failure Consumer Reports for Enhanced Service Efficiency,” IEEE 90th Vehicular Technology Conference (VTC 2019-Fall), 2019.
  • M. Amoozadeh, B. Ching, C-N. Chuah, D. Ghosal, and H. Michael Zhang," VENTOS: Vehicular Network Open Simulator with Hardware-in-the-Loop Support," 10th International Conference on Ambient Systems, Networks, and Technologies (ANT), April/May 2019.
  • C. Yen, D. Ghosal, M. Zhang, C-N. Chuah, and H. Chen," Falsified Data Attack on Backpressure-based Traffic Signal Control Algorithms," IEEE Vehicular Networking Conference (VNC), December 2018.
  • J. Wu, D. Ghosal, H. M. Zhang, C-N. Chuah, "Delay-based Traffic Signal Control for Throughput Optimality and Fairness at Isolated Intersection," IEEE Transactions on Vehicular Technology, vol. 67, no. 2, pp. 896-909, February 2018.
  • H. Chai, M. Zhang, D. Ghosal, C-N. Chuah, "Dynamic Traffic Routing in a Network with Adaptive Signal Control," Elsevier Transportation Research Part C, vol. 85, pp. 64-85, December 2017.
  • H. Chai, M. Zhang, D. Ghosal, and C-N. Chuah, "Dynamic Traffic Routing in a Network with Adaptive Signal Control," Transportation Research Board 2016 Annual Meeting, January 2016.
  • M. Amoozadeh, A. Raghumaru, C-N. Chuah, D. Ghosal, H. Michael Zhang, J. Rowe, and K. Levitt, "Security Vulnerabilities of Connected Vehicles Streams and their Impact on Cooperative Driving," IEEE Communications Magazine - Automotive Networking Series, 53(6), pp. 126-132, June 2015. [pdf]
  • M. Amoozadeh, H. Deng, C-N. Chuah, H. Michael Zhang, D. Ghosal, "Platoon Management with Cooperative Adaptive Cruise Control Enabled by VANET," Elsevier Vehicular Communications, 2(2), pp. 110-123, April 2015. [pdf] (VehCom Best Paper Award 2018)
Please refer to VMesh/VGrid page for our previous publications on vehicular networking.

Simulator

VENTOS is an integrated C++ simulator for studying vehicular traffic flows, collaborative driving, and interactions between vehicles and infrastructure through WAVE-enabled wireless communication capability.

Education Outreach

Our team involves several undergraduates in our research over the course of the project, including the offering of the first senior design project in self-driving car in 2017-18. This effort is supported via REU funding from NSF CMMI-1301496 grant and donations from Nvidia.

Funding

This work is supported by the National Science Foundation (NSF) Grant CMMI-1301496, NSF HDR:TRIPODS grant CCF-1934568, and Department of U.S. Department of Transportation's Center for Transportation, Environment, and Community Health (CTECH)