基于agent的自适应交通信号协同控制方法研究-电气工程专业论文.docxVIP

II II Agent-Based Adaptive and Cooperative Traffic Signal Control Abstract Current research about the optimization of intersection signal control mainly concentrated on signal timing with respect to single intersection problems. Only take single intersection problems into consideration reduces the complexity of traffic signal control. However, oversimplified issues make research results cannot be well applied into practice. In order to achieve better regional transportation control system results based on the collaboration of Agents，An Agent-based adaptive traffic signal control method was proposed to optimize traffic control and to improve vehicle driving behavior. Firstly, in order to obtain adaptive and cooperative traffic signal control，An Agent- based CVIS (Cooperative Vehicles-Infrastructure System, CVIS) was proposed to provide reliable real-time status data from the traffic flow. All cars and major intersection lights in the architecture are multi-agent based, including road-side units, car terminal and pedestrian detectors to achive people-vehicle-road and to perceive comprehensive information of the surrounding environment through the wireless communication network. Secondly, an adaptive traffic signal coordination control model ACTAM (Adaptive and Cooperative Traffic Light Agent Model) was created in the Agent-based CVIS. In ACTAM, the autonomy and self-organized creature of Multi-Agent was used to improve the adaptability of intersection traffic signal control. The Agent-based CVIS can obtain information from the traffic flow in real-time and optimize traffic signal control through reinforcement learning strategy. Thirdly, the use of Multi-Agent genetic algorithm (Multi-Agent genetic algorithm, MAGA) proposed a global optimization method for traffic signal control, thus junctions III III unknown signal control programs of other intersections can achieve adaptive and cooperative regional traffic control with each other. Finally，Combined with data provided