Fusion model of vehicle positioning with BP neural network

A vehicle positioning fusion model adopted Back-Propagation (BP) neural network is proposed in this paper, which presents a combination of Global Positioning System (GPS) and Mobile Positioning System (MPS) of lower cost and accuracy. The BP Algorithm is employed, and the problems of the slow convergence speed of the BP algorithm and the local minimal point can be solved utilizing the momentum method and the strategy of adaptive learning-rate. Training results with research data shows that this algorithm is applicable. The model is proved to be less depended on the positioning models of GPS and MPS and less cost consuming except for certain errors of position accuracy. Hence we also give result analysis for advanced ideas and improvements.