Maximum Correntropy Criterion Kalman Filter Based Target Tracking with State Constraints

Abstract

Since the maximum correntropy criterion (MCC) can capture higher moments of the stochastic signal, the maximum correntropy criterion Kalman filter (MCC-KF) behaves well in state estimation under non-gaussian noise compared to the classic KF. If some prior information such as state constraints is known, we can improve the performance of the MCC-KF by the incorporation of state equality constraints. In this paper, we realize this incorporation through estimation projection technique to deal with the linear and nonlinear constraints. The corresponding algorithm is presented as the main contribution of this paper. Applied in the example of target tracking, the algorithm shows its superiority against its classic counterpart.