Intelligent Control Of Signalized Intersections And Its Traffic Forecast Method

With the development of urban traffic and the wide use of the technologies of communication, computer and automation on transportation, people began to apply new theories and technologies on traffic control. This thesis is mainly focused on the application of Fuzzy and Neural Network theory on the signal timing technology of signalized intersections and the traffic flow forecast. The main works and research results consist of five parts: 1. Design a new fuzzy controller where there is great potential for traffic control with regard to high complex and stochastic characteristics. Vehicle delay is usually regarded as objective to decide optimal signal timing. However, in this fuzzy controller vehicle queue length replaced vehicle delay as performance index. The main structure of this controller is introduced in this thesis and traffic flow simulation software is used for giving results to show how successful this fuzzy controller. Different results of kinds of traffic states are also given by means of traffic simulation. 2. Based on research results all over the world those years, theory explanation is given to describe the simulation results of different signal timing methods. That will help us to improve urban traffic management in China. 3. The application of Genetic Algorithms(GAs) on transport volume forecast is proposed and a new point of view about how to build fitness function by means of sysmatic error is also proposed. 4. Traffic flow prediction model at signalized intersections based on fuzzy logic system and neural network is proposed. The most attractive strait of this model is that it is very practical because of its high accuracy and quick covergency. Present prediction methods are mainly focused on the application of neural network to improve the...