A new adaptive unscented Kalman filter based on covariance matching technique
Li Li, Changchun Hua, Hongjiu Yang · 2014
This paper develops a new adaptive unscented Kalman filter based on covariance-matching technique for nonlinear system with unknown statistical characteristics of the noises. The variances of process noise and measurement noise can be estimated online simultaneously by using information of innovation and residual sequence, and the accuracy of UKF will be improved. Furthermore, we show that the statistical convergence property and the stochastic stability of this technique. Numerical examples illustrate the effectiveness of the proposed adaptive filter design scheme.