The Research Of Intelligent Traffic Controller Based On Embedded System

At first,according to complex,random and indistinct characteristics of the traffic system,we can not established exact mathematic model. The fuzzy control method need not established exact mathematic model and artificial neural networks have parallel computing, distributed-memory and strongly learning capability. Combined with fuzzy control and neural network,this paper presented traffic real-time FNN (fuzzy neural networks) control method based on vehicles waiting queue in isolated intersection. A FNN controller was designed and matlab simulation was exploited to simulate system. The result showed that efficiency was satisfied.Considering the disadvantage of traffic signal controller that is used at present,The paper also introduces in detail a set of whole design Intelligent Traffic Signal System,including hardware and software. The main contents in this dissertation includes:with the consideration of traffic signal control in an four-phase intersection,integrated fuzzy control and neutral network was used,a FNN controller was designed. Matlab computer simulation showed that the result was very good;The hardware of traffic signal controller has been designed. It included the kernel control circuit,man-machine interface circuit,debugging interface module, power driver circuit as well as communication module based on Ethernet and USB. The design and development of embedded software has been elaborated in detail. Embedded software consisted of the building of cross-development embedded Linux building of root environment,the transplant of U-Boot and the development of Linux driver programs, the file system and JFFS2 file system,and the development of application software.