Comparative Study on Adaptive Fading Kalman Filter

The reasons of the instability of Kalman filter are analyzed from the stability of Kalman filter.A new approach to adaptive estimation of Kalman filter fading factor is developed and is compared with the strong tracking Kalman filter.The characteristic that the filter residuals are zero 2 mean Gaussian white noise vectors is used and a chi 2 square distribution variable is made while computing the fading factor.Simulation result shows that the proposed method has the ability of restraining filtering divergence under the condition of wrong system noise attributes and has better efftect of estimation.The derivation of new method is simple and the computation burden is low enough to adapt to calculate online.