An Improved Fading Kalman Filter and its Application to GPS Kinematic Positioning

Abstract:

Based on mediumlong baseline GPS kinematic positioning model and the theory of classic fading Kalman filter,a modified fading factor vector is proposed.Different widths of sliding windows are used to obtain a series of parallel filters.The weight factor of covariance is decided by the residuals of every filter.The modified fading Kalman filter is more effective and sensitive than classic Kalman filter.Various experiments are made to test the performance of this method.The results show the precision of longbaseline GPS kinematic positioning is improved.

Key words:GPS kinematic positioning; fading Kalman filter; fading factor vector; sliding window