DRL-TP3: A learning and control framework for signalized intersections with mixed connected automated traffic

The emerging connected and automated vehicle (CAV) technologies offer new opportunities for urban signalized intersection management. Through wireless communication and advanced sensing capabilities, CAVs can detect the surrounding traffic environment and share real-time vehicular information with each other and the infrastructure, and individual trajectories of CAVs can be precisely controlled. This paper proposes a real-time learning and control framework for signalized intersection management, which includes both signal optimization and CAV trajectory control. The proposed framework integrates perception, prediction, planning, and optimization components and aims at improving efficiency mixed connected automated traffic in terms of traffic throughput and delay. This framework applies the Long Short Term Memory (LSTM) networks to implicitly learn traffic patterns and driver behavior and then estimate and predict the microscopic traffic conditions that are only partially observable. Then it utilizes deep reinforcement learning (DRL) to solve signal optimization problems by learning from the dynamic interactions between vehicles and the traffic environment. Under the proposed framework, the vehicular trajectories of CAVs can be controlled to maximize the utilization of the green time and reduce the start-up lost time by using a highly efficient trajectory planning algorithm. The CAV platooning operation, in coordination with traffic signals, is also implemented such that CAVs can pass the intersection efficiently. Simulations are performed at a signalized intersection a signalized intersection with multi-lane approaches, high traffic demand, and standard ring-barrier control, and results show that the proposed DRL-TP3 framework can significantly improve the throughput and reduce the average delay across different CAV market penetration rates (MPRs). The authors also investigate the impacts of different sensor capabilities of unobservable vehicle estimation and implementation of a lane change prohibition zone under the DRL-TP3 framework.

• Record URL:
  https://doi.org/10.1016/j.trc.2021.103416
• Record URL:
  http://www.sciencedirect.com/science/article/pii/S0968090X21004101
• Availability:
  • Find a library where document is available. Order URL: http://worldcat.org/issn/0968090X
• Supplemental Notes:
  • © 2021 Elsevier Ltd. All rights reserved. Abstract reprinted with permission of Elsevier.
• Authors:
  • Guo, Yi
  • Ma, Jiaqi
• Publication Date: 2021-11

Language

• English

Media Info

• Media Type: Web
• Features: Appendices; Figures; References; Tables;
• Serial:
  • Transportation Research Part C: Emerging Technologies
  • Volume: 132
  • Issue Number: 0
  • Publisher: Elsevier
  • ISSN: 0968-090X
  • Serial URL: http://www.sciencedirect.com/science/journal/0968090X

Subject/Index Terms

• TRT Terms: Connected vehicles; Human beings; Machine learning; Motor vehicles; Signalized intersections; Traffic platooning; Trajectory control; Vehicle mix
• Subject Areas: Highways; Operations and Traffic Management; Vehicles and Equipment;

Filing Info

• Accession Number: 01788450
• Record Type: Publication
• Files: TRIS
• Created Date: Nov 17 2021 2:27PM