Mixed Microscopic Traffic Simulation Based On Cellular Automata And Multiagent Research

Traffic problems are global hot issues, and the traffic simulation technology has become an important means to research and resolve traffic problem. Different from the traffic status aboard, the traffic flow in our country has many unique characters. And it cannot be simulated truly by most of traffic simulation software. To solve this problem, this thesis puts forward a new research means which is based on the combination of Cellular Automata and Multi-Agent System (MAS) theories. And on this base, a mixed microscopic traffic simulation system is proposed which includes motor vehicles, bicycles and pedestrians.Firstly, with research on the theory of Cellular Automata and Multi-Agent System, this thesis analyzes their advantages and disadvantages in traffic simulation applications, demonstrates the feasibility of combining these two theories. Then, based on some skillful extensions and modifications of the two theories, this thesis finds out the combination points and methods with the two theories.On this base, a mixed microscopic traffic simulation system is built. The system sets up a architecture based on CA theory, and brings forward a new thought to divide the system into two layers of traffic entities layer and signal control layer, the elements in the two layers influence each other by mapping . In the traffic entities layer, by the means of divide a lane to many cells, a single road section is built which is composed of many lanes, and then the road network of the whole system is composed of many road sections. And the road section includes the straight and the crossing. Before entering the crossing, the moving estates of the entities are influenced by the traffic light mapping from the signal layer. In the crossing, the travel states and tracks of the entities is decided on the priorities of the roads and the destinations.In this thesis, by some skillful extensions and modifications of the cell, some cells have the characters of Agents. And a Multi-Agent System is built which consists of many traffic entity Agents. Through research the moving characteristics of the motor vehicles; the thesis establishes a vehicle Agent model including vehicle generation model and vehicle action model. The vehicle action model includes the most important driving behaviors including free-driving, car-following and lane-changing. Based on the research of the moving characteristics of bicycles and...