| With the increase of vehicle ownership, traffic flow is increasing which increases the pressure of intersection. However, most of the traffic lights are controlled by fixed–time scheme, which can not make the most use of intersection. So a rational intelligent traffic light control system should be established.Based on this background, firstly, this paper analyzes the key technology and development for intelligent traffic light control system; secondly establish a intelligent traffic light control system based on image processing, which contains two modules: one is image processing module and another is intelligent control module; finally, illustrate its effectiveness by simulation experiment.In the image processing module, a method based on liner equation to extract drive line is proposed; a rapid method to extract single lane is proposed which is based on temple and drive line; a fast background extraction based on single lane is proposed to improve the real time; along the road length and width, a formula of the actual length represented by row coordinate is created which can be used to measure the length of vehicle queue and recognize vehicle and get the equivalent length of vehicle queue for each phase.In the intelligent control module, integrating the merits of both fuzzy control and neural network control, fuzzy neural network control is proposed, in this system the expertise can be fully made use of and it is also capable of automatism learning; developing simple self-learning rules, so the system can automatically obtain samples and complete their own training which can make the system to adapt changing traffic conditions.In this paper, with the standard of average delay time, simulation experiment about fixed-time control and fuzzy control and fuzzy neural control has been done. The experiment shows that fuzzy neural control has excellent control performance when traffic flow is moderate or uneven and shows that it can be adaptable to traffic flow change... |
PDF Full Text Request