Multi-information Fusion and Filter Study of Multi-sensor Velocity Measurement on High-speed Train

摘要

For the rapid development high-speed railway system, improvement approach of the velocity measurement accuracy has been studied based on multiple speed sensors on high-speed train. In this method, the velocity measurement data from multi-channel speed sensors were dealt through data fusion of arithmetic mean filter, weighted arithmetic mean filter, Federated Kalman filter and adaptive Federated Kalman filter algorithm. On this basis, the comparative study was carried out both at high speed and at low speed based on weighted average algorithm, and algorithm of Federated Kalman filter and adaptive Federated Kalman filter were designed. Discussing the adaptive Federated Kalman filtering problem that four channel sensors are normal and one of sensors is faulted. Then simulation parameters and coefficients were set according to the algorithm and simulated in MATLAB. The results show that it can achieve better fusion effect base on Federated Kalman filter and adaptive Federated Kalman filter algorithm. And the adaptive Federated Kalman filter algorithm is applied to high-speed train system, which has improved the velocity measurement accuracy and fault tolerance, and made the high-speed railway system has better adaptability and improve the train's operating efficiency based on controlling trains safely running.