Sage windowing and random weighting adaptive filtering method for kinematic model error

This paper presents a new method for adaptive estimation of kinematic model error in dynamic aircraft navigation. This method combines the concepts of random weighting and Sage windowing to online monitor predicted and observation residuals to control the influence of the kinematic model's systematic error on system state estimation. Based on the Sage windowing, random weighting estimations are constructed within a moving time window for the systematic error of the kinematic model as well as the covariance matrices of the observation noise vector, the predicted residual vector, and the predicted state vector. Experimental results and comparison analysis demonstrate that the proposed method not only adjusts the covariance matrices of the observation noise vector and the predicted residual vector, but also effectively controls the influence of the kinematic model error on state parameter estimation, thus improving the navigation accuracy.