Research on cooperative control of vehicle speed and traffic signal based on V2X environment

Abstract

Based on the environmental information provided by V2X (Vehicle-to-Everything) technology, the collection of vehicle position information and motion state information, combined with intersection signal information, can improve the traffic efficiency of signalized intersections on urban roads. This paper discusses a cooperative control method of vehicle speed and traffic signals for intelligent connected vehicles, which offers insights for the traffic management system. In this paper, the motion state of vehicles at signalized intersections is comprehensively considered by employing the partition control concept. Based on the deep reinforcement learning algorithm known as DQN (Deep Q-Network), the environmental states of the vehicle, including the position matrix and speed matrix, are constructed. The action space is designed to account for the signal phase in relation to the traffic congestion length at the intersection. Additionally, the convergence of the control model is enhanced by incorporating the greedy algorithm and the experience replay mechanism. Finally, a joint simulation using SUMO-Python is designed to verify the effectiveness of the method proposed in this paper. The results show that compared with traditional timed signal control and single-vehicle speed control, the speed and traffic signal cooperative control method proposed in this paper can effectively enhance traffic and environmental benefits.