A Robust Adaptive Traffic Signal Control Algorithm Using Q-Learning under Mixed Traffic Flow

The operational and safety performance of intersections is the key to ensuring efficient operation urban traffic. With development automated driving technologies, ability adaptive traffic signal control has been improved according data detected by connected vehicles (CAVs). In this paper, an was proposed optimize intersection. algorithm based on Q-learning considers loop detectors CAVs. Furthermore, a comprehensive analysis conducted verify algorithm. results show that average delay conflict rate have significantly optimized compared with fixed timing actuated control. addition, good in test irregular provides new idea for intelligent management isolated under condition mixed flow. It research basis collaborative multiple intersections.