Real-Time Adaptive Traffic Signal Control in a Connected and Automated Vehicle Environment: Optimisation of Signal...

Area of Science:

• Transportation Engineering
• Artificial Intelligence
• Urban Planning

Background:

• Adaptive traffic signal control (ATSC) aims to minimize urban traffic congestion and delays.
• Reinforcement learning (RL) shows promise for optimizing traffic signal plans.
• Integrating RL with connected and automated vehicles (CAVs) presents an open challenge.

Purpose of the Study:

• Develop a real-time RL-based ATSC system for urban traffic management.
• Optimize signal plans to minimize total vehicle queue length.
• Decrease total vehicle stop delays by integrating speed guidance for CAVs.

Main Methods:

• Implemented a real-time reinforcement learning (RL) algorithm for adaptive traffic signal control.
• Combined RL-based signal optimization with a speed guidance system for connected and automated vehicles (CAVs).
• Evaluated performance using two measures: total queue length and total stop delays.

Main Results:

• The proposed RL-based ATSC with speed guidance significantly outperforms fixed timing plans and traditional actuated control.
• Demonstrated reductions in average vehicle stop delay and queue length.
• Effectiveness is particularly pronounced under saturated and oversaturated traffic conditions.

Conclusions:

• The integration of RL-based traffic signal control and CAV speed guidance offers a superior approach to traffic management.
• This hybrid system effectively mitigates urban traffic congestion and improves network performance.
• The method shows significant potential for future intelligent transportation systems.