Intelligent cooperation control of urban traffic networks

Abstract

Area-wide traffic control in urban areas is complicated. Traffic conditions are often changeable and random. This paper develops and tests a distributed system of urban traffic control, for use within traffic signal controlled networks. The system uses multi-agent and information fusion technology. The system framework and agent structure are discussed in the paper. The intelligent cooperation algorithm used by the proposed urban traffic control system is based on genetic reinforcement learning. The system is capable of responding to changing traffic conditions in real-time. It has the ability to coordinate and optimize the traffic signal controlled network through system self-study and adaptability. The validity of the algorithm is tested using simulation analysis.