Adaptive Kalman filter based on multiple fading factors

A scalar fading factor is calculated in the existing adaptive fading Kalman filter and each filtering channel just gets the same adjustments,which is unfavorable for improving the filtering accuracy.Aimed at this issue,a new method based on the estimate covariance and the innovation covariance estimator is proposed.A set of innovation covariance estimators based on the limited memory index weighted method works in parallel to calculate the fading factors and then the factors are distributed to each filtering channel according to the estimate covariance to improve the performance of the adaptive Kalman filter.Simulation results show the effectiveness of the proposed method.