An Information Fusion Algorithm for Multiple GPS Track Data

Accurate GPS (Global Position System) electronic map is the basis of the train positioning and control using GPS as the main positioning manner. As high precision is contradictive to low cost, a new measurement method with high efficiency and low cost is proposed in this paper to get the GPS data with high accuracy for railway track. A heuristic algorithm based on continual split and non-linear optimization is brought forward to realize the information fusion of multiple GPS track data with fixed endpoints. The results of the experiments show that the algorithm can learn the high-accuracy GPS data from the multiple low-accuracy GPS data. The algorithm´s accuracy gets better with the increase of line segments.