Iterative Tuning With Reactive Compensation for Urban Traffic Signal Control

作者：Wang, Yu^*; Wang, Danwei; Jin, Shangtai; Xiao, Nan; Li, Yitong; Frazzoli, Emilio

来源：IEEE Transactions on Control Systems Technology, 2017, 25(6): 2047-2059.

DOI：10.1109/TCST.2016.2634459

• Iterative tuning (IT)
• preemptive and predictive control
• traffic flow patterns
• urban traffic signal control.

To utilize the utmost capacities of junctions with traffic signals, having preemptive and predictive capabilities with integrated traffic sensors are important elements for an intelligent traffic signal control system. Recently, an efficient and preemptive traffic signal control strategy known as iterative tuning (IT) has been developed by exploring repetitive characteristic of traffic demand on working days, whereas on weekends or days with special events, severe nonrepetitive disturbances limit the performance of IT strategy practically. This paper proposes a preemptive and predictive methodology, i.e., IT strategy with reactive compensation (ITRC), to reduce network delay time and eliminate unnecessary stops for vehicles. With anticipation of repetitive traffic flows based on historical data, nominal traffic signals are tuned by IT controller over repetitions. The reactive compensation, which is junction-based model predictive control strategy (JMPC), makes adjustments on nominal traffic signals and compensates the nonrepetitive elements. Rigorous analysis provides sufficient condition for guaranteeing the convergence of ITRC. Case study of Commonwealth Avenue in Singapore with heavy traffic from morning peak to evening peak hours is presented, and the simulation results show the effectiveness, robustness, and convergence of ITRC over pure IT strategy and pure JMPC strategy for scenarios with various traffic variations, respectively.

• 出版日期2017-11
• 单位南阳理工学院; 北京交通大学; MIT