Dynamic Optimal Path Planning Based On User Selection Under Multi-restrictions

With the development of economic and technological rapidly,cars are becoming an indispensable tool in people's lives,making people travel more and more convenient and expand the scope of activities of the people.Because travelers are strange to the traffic conditions and geographical environment,travelers may get lost or run into traffic jams,so that travelers have the correct travel route is particularly important.Traditional path planning only consider shortest-distance or shortest-time,and ignore user's need and real-varying traffic information.In this paper,planning the optimal path dynamically based on user preferences and real-time traffic information,and allowing users to travel more secure and convenient.Firstly,analyzed the traffic state with a strong time-variable,time and space characteristics.In order to make users know the accurate traffic state,presented the model that predicts the traffic state and changing space based on time series.Constructing the wavelet neural network from time dimension and spatial dimension to predict traffic state,fusing the result of prediction through the fusion algorithm to improve the prediction accuracy.Meanwhile,predicting the traffic state changing space by the fuzzy information granulation SVM model,fuzzy information granulation fuzzy the historical data to get the minimum,average,and maximum of traffic state,and using Support Vector Machine to predict them to know the changing space of road traffic state.Allowing users to know the correct real-time traffic information and rationally adjust travel routes.Secondly,analyzed the representation methods of road network,weighted graph is abstracted from actual road network by road section-chain model to solve the road restrictions.In order to make the path planning to be accord with users' preference and real-time traffic information,multi-objective road impedance function model is proposed.Analysis the influence of the vehicle traveling,give full consideration to the users' preference,combined with dynamic factor and static factor,using subjective weight determining method and objective weight determining method to determine the weight of each factor,calculate the value of comprehensive road impedance by multi-objective road impedance function model.Finally,optimized the traditional Dijkstra algorithm from three aspects: data storage structure,road network structure division and path search mechanism,improved the efficient of the algorithm.Research of the part road network of Chongqing,combined with the traffic state prediction model and multi-objective road impedance function model previously,determine the comprehensive impedance of each road section,then,calculate the optimal path by Dijkstra algorithm.After the example verification,the proposed model can plan and adjust optimal path dynamically based on user's preferences and real-time traffic information all the time,reduce the cost of users' travel and remit the traffic jam.