Multimodel Train Speed Estimation Based on High-Order Kalman Filter

Accurate estimation of train operation state is very important to ensure the safe and efficient operation of high-speed trains. Train running speed, as a key link, directly affects the train travel time, energy consumption, and cooperative operation with other trains. In this article, a train multimode model considering the actual operating environment is established, and a novel train speed estimation method based on high-order Kalman filter is proposed. In this method, an auxiliary model is introduced, the solution of the statistical characteristics of the auxiliary model errors is given, and the high-order information in the nonlinear model is captured and used. The accurate estimation of train running speed is realized by improving the prediction accuracy of train running speed. The effectiveness of the proposed method is verified by different motion modes of CRH3 high-speed train with four motion and four tows. The simulation experiments show that compared with the high-order filtering methods such as IEKF and HOUSE, the estimation accuracy of the proposed high-order Kalman filter can be improved by 51.7% and 98.45%, respectively.