An adaptive Kalman filter for dynamic estimation of harmonic signals

In electrical railway systems there is often a need of detecting or/and predicting harmonic signals contained in measurement data for vehicle control or monitoring purpose. An efficient on-line estimation method for such applications is the Kalman filter technique. However, the performance of a standard recursive Kalman algorithm is strongly dependent on the a priori information of the process and measurement noise which is either unknown or is known only approximately in practical situations. Furthermore, a Kalman filter often suffers from the problem of “dropping off” and loses then the ability to match abrupt parameter changes. In this paper an adaptive Kalman filter based on correlation analysis is proposed to help overcome these problems. The modelling and estimation technique is described in the paper. Simulation results using measured vehicle line current demonstrate the effectiveness of the proposed method