Study on Signal Light Control of BP Neural Network Intersection Based on Swarm Intelligence Algorithms
Download as PDF
With the development of urbanization and the popularization of private cars in China, urban traffic congestion has become increasingly prominent. How to solve this problem through practical, effective and feasible methods has attracted wide attention. In this paper, the structure of signal control system based on reinforcement learning of BP neural network based on swarm intelligence algorithm is proposed. By applying reinforcement learning, the optimal control problem of signal lamp is transformed into the decision-making problem of switching operation phase. A coordinated control method of regional signal based on BP neural network is designed. The method preserves the neural network self-learning structure and considers the number of vehicles arriving between adjacent intersections in the region. The simulation results verify the effectiveness of the BP neural network based regional signal coordination control method. The system can sense the change of traffic flow and adaptively adjust the signal switching strategy to achieve the optimal control effect. This method is feasible and has obvious advantages compared with timing control.
BP Neural Network, Intersection Signal Lamp, Fuzzy Control