Resrarch Of Adaptive Cooperative Positioning Technology For Vehicular Network

With the rapid development of urban transportation,traffic problems are increasingly prominent and gradually becoming vital livelihood issues,while emerging intelligent transportation technology such as vehicle networking technology has become a new developing trend.Vehicle positioning technology is a basic and important part of vehicle networking technology,especially in some scenarios involving safety which need lane level high-precision positioning.At present,GPS is commonly used to provide positioning information for vehicles,but the GPS equipment currently used in many vehicles has limited accuracy,and in urban environments,due to the occlusion of tall buildings and other obstacles,GPS often has serious faults.Currently,the development direction of vehicle networking positioning technology mainly includes inertial navigation,differential positioning,sensor-assisted positioning and cooperative positioning,while this paper focuses on the research of cooperative positioning technology in vehicle networking.Cooperative positioning technology optimizes and integrates the status information and physical layer information such as signal strength and carrier frequency shift through mutual communication between vehicles,in order to obtain more accurate location information in the case of limited data,in which coupled cooperative positioning technology has better performance.Consequently,how to use limited data sources to improve vehicle positioning accuracy as much as possible is of great significance.Aiming at the low GPS positioning accuracy of most vehicles and the urban environment where GPS signals are degraded or denied.this paper proposes a cooperative positioning algorithm based on the information of GPS,RSS and CFO.The simulation results show that the cooperative positioning algorithm can improve the positioning accuracy by 59% compared with the original GPS signal when GPS is not severely restricted.Moreover,the advantages of the cooperative localization algorithm based on GPS/RSS/CFO are illustrated by comparing with other coupling algorithms.And this algorithm is equally effective in severely constrained scenarios that can occur in urban environments.Furthermore,on the basis of the above algorithm,the paper continues to make targeted improvements.On the one hand,through the analysis of RSS technology,a distributed dynamic decision function based on CRLB is proposed,through which the decision center can make decisions adaptively according to the actual situation.On the other hand,aiming at the inaccurate model of GPS Kalman filter system,a covariance matrix correction algorithm based on BP neural network is proposed,which can dynamically adjust the covariance matrix of Kalman filter according to the running state of the vehicle.Through simulation analysis,the improved cooperative localization algorithm has better performance than the original algorithm and other similar algorithms,and can basically meet the requirements of lane level high-precision positioning when GPS is not severely restricted.