The improved Kalman filter algorithm based on curve fitting

In order to improve the effect of tracking dynamic object, an improved Kalman filter algorithm based on curve fitting is given. When the target is maneuvering, the system model of Kalman filter cannot match exactly, filtering accuracy will reduce or even diverge. Therefore, credibility of state predictive value in the filter decline, and filtering should depend more on measuring value. Curve fitting based on historical trace reflects maneuvering information. Curve fitting combined with the Kalman filter, better describes the target mobile. Monte Carlo simulations showed that the improved algorithm have better accuracy than conventional Kalman algorithm and keep the characteristic of structure simple and small storage.