Study On Fusion Method Of High-speed Train Running Information

High-speed train running information is the foundation of real-time dynamic tracing monitoring and decision-making behavior, it have the characteristics of variety and complexity. Although the development of sensor technology to provide information for the security and convenience, but the breadth and depth in the access to information remain s to be further improved, and the real time and accuracy of information need to be further improved. Using information fusion method can effectively improve the integrity and the accuracy of information and reliability of information transmission.This paper based on CTCS-3 trains to study train running information fusion method.It looked the train speed information detection as a state estimation problem, the problem of train positioning as the target tracking problem, studied mainly the train speed information fusion method, the train position information fusion method based on the "current" statistical model,the train intelligent decision information fusion method,the simulation results show that the information fusion method is adopted to train running information processing, It has good effect to improve the accuracy and reliability of the information.In the research of CTCS-3 train operation control system, first of all, from the train speed information acquisition modes and system requirement, the velocity information fusion scheme is determined. Considering it is not accurate for dynamic mathematical model and noise model of high-speed train, the speed information for dynamic state is estimated using extended kalman filtering method, to get the optimal estimation of each road speed sensor measurement information, reused the main filter for filter results fusion processing to improve the speed of decision-making, restrain the filtering divergence. The simulation results show that the speed of the fused information is more accurate.From the Angle of target track fusion train locating method are studied, by using the "current" statistical model, the combination of GPS/DR integration positioning, train positioning fusion model is set up, and put forward adopting Reset type(No- Reset, NR) fusion location and federated kalman filter, the simulation results show that has good tracking performance.For the decision object of train operation control system, the characteristics of complicated process is the decision results are difficult to determine, this paper studied the fuzzy reasoning information fusion technology, established the model of information fusion of fuzzy reasoning, said the result of train control expert knowledge with SPN method, simulation shows that use the train operation control system combined with fuzzy reasoning information fusion technology to make decisions, has good control effect.