Numerically stable minimum error entropy Kalman filter

Recently, the Kalman filter algorithm based on the minimum error entropy criterion has been presented to outperform the conventional Kalman filter in the case of impulsive noise. In practice, it may be unstable in numerical calculation. This paper proposes a robust algorithm to solve the problem of instability in numerical calculation. The convergence and stability of the algorithm are verified by the performance analysis and simulations. In the mean square behavior analysis, we propose a method to estimate the steady-state errors. Simulations show that the experimental steady-state errors of the algorithms agreed with the theoretical values. (C) 2020 Elsevier B.V. All rights reserved.

Automated summary

A robust algorithm has been proposed in this paper to address the instability in numerical calculation of the Kalman filter algorithm based on the minimum error entropy criterion when dealing with impulsive noise. The convergence and stability of the algorithm are verified through performance analysis and simulations, and a method to estimate steady-state errors is proposed in mean square behavior analysis. Simulation results show that the experimental steady-state errors of the algorithm match the theoretical values.