Research On Vehicle Location Algorithm And Map Matching Algorithm Based On Internet Of Vehicles

In the current intelligent transportation field,people’s demand for high-precision vehicle positioning and accurate road network matching is increasing.The research of lane level matching algorithm is of great significance for the safe running of vehicles and the improvement of traffic efficiency.However,when the satellite signal is interfered,the vehicle positioning will appear positioning drift,positioning error increases and other conditions,and lane level matching is very dependent on high-precision map,the cost is expensive,and can not be widely circulated in the user market for the time being.Based on the above factors,this paper realizes vehicle positioning through the fingerprint positioning method in the environment of the Internet of vehicles,improves the positioning accuracy,and proposes a lane level map matching algorithm,so that vehicles can achieve lane level map matching under the condition of ordinary electronic map.The main research contents are as follows:The principle of fingerprint location can be used to locate vehicles.An improved fingerprint database location matching algorithm based on BP neural network location using particle swarm optimization algorithm and genetic algorithm in the environment of vehicle network is proposed,which achieves the location accuracy of lane level.In order to solve the problem of insufficient parameters of the existing hidden Markov model in road matching,the algorithm of road matching was improved,and two variables of real-time vehicle speed and travel inclination Angle were introduced,and a hidden Markov model based on high order Gaussian function was proposed.On the basis of road matching,a lane matching method based on RANSAC algorithm and Freiche distance is proposed.This method can achieve accurate lane matching without the need of high-precision map,only the need of vehicle information,lane width and number.As for the scheme proposed above,this paper verifies through simulation,actual test and other ways,and the positioning accuracy reaches the average error of 0.28 m and the RMSE of0.22 m.Lane level matching algorithm can also achieve 96.9% matching accuracy.