Traffic Signal Control Optimization Based on Deep Reinforcement Learning with Attention Mechanisms

Deep reinforcement learning (DRL) methodology with traffic control systems plays a vital role in adaptive traffic signal controls. However, previous studies have frequently disregarded the significance of vehicles near intersections, which typically involve higher decision-making requirements and safety considerations. To overcome this challenge, this paper presents a novel DRL-based method for traffic signal control, which incorporates an attention mechanism into the Dueling Double Deep Q Network (D3QN) framework. This approach emphasizes the priority of vehicles near intersections by assigning them higher weights and more attention. Moreover, the state design incorporates signal light statuses to facilitate a more comprehensive understanding of the current traffic environment. Furthermore, the model’s performance is enhanced through the utilization of Double DQN and Dueling DQN techniques. The experimental findings demonstrate the superior efficacy of the proposed method in critical metrics such as vehicle waiting time, queue length, and the number of halted vehicles when compared to D3QN, traditional DQN, and fixed timing strategies.