Robust adaptive filter allowing systematic model errors for transfer alignment

This paper presents a new robust adaptive filtering method for transfer alignment by taking into account the systematic errors of observation and kinematic models in the filtering process. The proposed method overcomes the limitation of the traditional Kalman filter, that is, the requirement of precise kinematic and observation models for transfer alignment. It adaptively adjusts and updates the prior information through the equivalent weighting matrix and adaptive factor to resist the disturbances of systematic model errors on system state estimation, thus improving the accuracy of state parameter estimation. Experimental results and comparison analysis demonstrate that the proposed robust adaptive filtering method can effectively improve the performance of transfer alignment, and the achieved performance is much higher than those of the Kalman and traditional robust adaptive Kalman filters.