Extended Information Filter under Maximum Correntropy Criterion

This study concerns the information filtering problem for nonlinear non-Gaussian systems. Under the maximum correntropy criterion (MCC), in the information filter framework, a novel information filter named maximum correntropy extended information filter (MCEIF) is proposed for nonlinear non-Gaussian systems. Similar to the classical extended information filter (ElF) under MMSE criterion, the prior state estimate and prediction information matrix are calculated. Yet the estimation information matrix and filter gain information matrix are reconstructed by utilizing the MCC, and then the posterior state estimation and filtering information matrix are updated. The final numerical simulation results demonstrate the effectiveness of MCEIF.