Aiming at the problem of improving the efficiency of urban intersection control, two signal control strategies based on reinforcement learning (Q-learning (QL) and deep Q-learning network (DQN)), respectively, re proposed. Overcoming the rough and passive defects of the traditional intersection timing control, the QL and DQN algorithm with intelligent real-time control is adopted. An algorithm framework with radar and video detector data as input and optimal intersection control strategy as output is constructed. Based on the traffic simulation platform, a typical urban intersection is simulated and the control effect is tested. The results show that the proposed two intelligent control strategies can actively respond to different traffic states, converge in short training time and find the optimal control strategy. QL-based control strategy and DQN-based control strategy can effectively reduce the travel time by more than 20% and the stop delay by more than 30%. And DQN-based control strategy is more effective than the QL-based control strategy.

CICTP