Design And Realization Of Intelligent Traffic Path Planning System Based On Fuzzy Neural Network

Along with the deepening of urban transportation infrastructure construction and the increase of urban vehicles,urban road planning and vehicle routing has become an important factor affecting people's daily travel.As an important hotspot in intelligent transportation,computer application,network analysis and applied mathematics,the problem of urban vehicle routing has become the top task of urban road planning.At present,the traditional urban path planning can not deal with the massive non-linear data produced by the real-time traffic,and the information of different traffic management departments can not be interconnected,which undoubtedly further aggravates the difficulty of urban path planning.Therefore,it is very important to develop a path planning method with intelligent learning ability,which can carry on the urban vehicle path planning under the condition of real-time dynamic road network.This paper starts with the existing problems of intelligent transportation path planning in Nanjing,Jiangsu Province.Firstly,it studies the general path planning problem,analyzes the modeling of urban road dynamic network.Secondly,based on the road condition of Nanjing,a path planning algorithm based on fuzzy neural network inference system is proposed,and the related path risk avoidance mechanism is presented and the steps of the algorithm are given.Finally,based on the electronic map technology and VS2010 software development platform,prototype system is developed and field tests are conducted in Nanjing.The main work of this paper is as follows:(1)The feasibility of the current mainstream path planning scheme is studied,and the real-time dynamic road network modeling is studied deeply.This paper analyzes the characteristics of the traditional urban path planning scheme and expounds its shortcomings under the current real-time road network.It also studies the dynamic real-time road network modeling,including the road network connectivity method,the storage structure and the like.(2)This paper proposes a multi-path planning algorithm based on fuzzy neural network inference system and expounds its related path risk aversion mechanism.First,according to the characteristics of traffic conditions in Nanjing of Jiangsu Province,three types of path risk assessment input parameters are defined: vehicle driving operation loss,road control loss and noise control loss.Secondly layer neural network architecture is constructed according to the above path risk parameters.After the sample training on the fuzzy neural network inference system with the annealing algorithm,the layered training result is achieved.Finally,after the road network risk function output value is obtained from the fuzzy neural network inference system,the modified Clark-Wright algorithm is used to develop the multi-path planning of the vehicles under the road network.The results of the multi-path planning and the corresponding conclusions of the non-risk avoidance and risk aversion algorithms are given respectively.(3)Design and realization of intelligent transportation path planning system based on fuzzy neural network.Based on the multi-path planning algorithm with risk-aversion function,MAPX space of electronic map and VS2010-related software development framework,the algorithm is implemented in C ++ language,and an intelligent traffic path planning system based on fuzzy neural network is constructed.Field tests of the system is conducted,and the similarities and differences between the system adopting the risk aversion mechanism and the multi-path planning without adopting the risk aversion mechanism is analyzed.