Stochastic Stability of the Improved Maximum Correntropy Kalman Filter Against Non-Gaussian Noises

In this paper, an improved maximum correntropy Kalman filter (IMCKF) algorithm is proposed to enhance the estimation accuracy of conventional correntropy based Kalman filter against the non-Gaussian noise. To increase the proposed algorithm estimation precision, a novel cost function is introduced based on weighted factors. Then the IMCKF algorithm is put forward and derived in detail. Furthermore, the stochastic boundness of the estimation error is discussed to illustrate the IMCKF algorithm’s stability. Finally, simulation results demonstrate that the proposed IMCKF algorithm increases the estimation precision and robustness performance in contrast to the conventional Gaussian Sum Kalman filter and maximum correntropy Kalman filter.