Study on Information Fusion Algorithm for Multiple GPS Railway Track Data
Dewang Chen · Journal of the China Railway Society · 2011
To generate the GPS(Global Positioning System) electronic map for train control systems,GPS data with high accuracy are required.However,GPS receivers with high-accuracy cost too much.Therefore,learning high-accuracy GPS data from multiple low-accuracy GPS data measured by the low-cost GPS receiver through the information fusion algorithm is of great use.The existing information fusion algorithm can reach the target,but it is only applicable to simple tracks and the optimization time is too long.Based on the theory of principal curves and the special feature of the fixed high-accuracy points of interest in railways,we proposed the improved algorithm by adding a point in the area whose error is maximal and using local optimization in this paper.The results of the experiments show that the new algorithm can learn high-accuracy GPS data from multiple low-accuracy GPS data and has great advantages in the time of computation and the scope of application.