Algorithm And Simulation Of Signal Control Of City Single Intersection

As social and economic development, various countries pay a increasing attention to transportation issues. Traditional control methods have not effectively solved the increasingly serious traffic problems, and as the development of computers, more and more intelligent control technology is applied to the urban traffic control.In this paper, a single intersection signal control of the city is the research object; first descript a four phases, eight-lane isolated intersection traffic simulation model, study of urban traffic signal control by both on the theoretical algorithm analysis and system simulation achievement.Research is divided into the following three specific aspects in the terms of theoretical analysis on algorithm.①Fuzzy controller is designed by use of MATLAB based on the fuzzy control. Design procedure consists of fuzzification on input and output, determination on membership function, determination on the fuzzy rules and defuzzification, then get fuzzy control timing scheme;②Given that fuzzy control don't have self-learning ability so that it can analyze characteristics of neural network, the paper study on the design of fuzzy BP neural network controller by use of fuzzy control combined with neural network, which consists of design on the fuzzy BP neural network structure and BP neural network, network test by use of MATLAB, fuzzification on input, clarity on output, and then get the fuzzy BP neural network control timing scheme;③BP neural network is static network while Elman neural network is dynamic network. In the foundation of the study on the fuzzy BP neural network control, make tentative control on Single intersection signal by use of fuzzy Elman neural network control, correspond fuzzy Elman neural network controller is designed. Design procedures consists of design of Elman network, network training and testing in MATLAB, fuzzification and clarity of network, and then get fuzzy Elman neural network control timing scheme.On system simulation, in order to compare the timing plans achieved by the above three control algorithms, the paper design urban single intersection signal control simulation system with object-oriented C++ programming language. System includes three modules of the basic system operation module, the signal control module and the data statistics module. Reproduce the comparatively real traffic scene though real-time dynamic visual interface, and observe the actual car congestion in a single intersection case in different traffic conditions and different control mode, and through the statistics as a supplementary tool compare the control effect of several control methods. Finally, in order to compare the effect of timing scheme with the traditional timing control, also realize the timing control simulation.Simulation results show that compared with fuzzy control timing control effectively reduce the average vehicle delay, compared with fuzzy control, fuzzy neural network control has further improved performance. In the case of high traffic, the average delay of Fuzzy Elman neural network control is lower than BP neural network fuzzy control, in low traffic situations, the situation is opposite.