A novel robust interacting multiple model filter for manoeuvering target tracking

Abstract

AbstractA robust interacting multiple model approach is proposed to address the problem of accuracy and non‐Gaussian measurement noise in manoeuvering target tracking. The proposed approach introduces multiple fading factors into the prediction covariance matrix and adjusts each channel of the gain matrix in real time to improve the accuracy caused by model mismatch and enhance state transition capability. Considering the non‐Gaussian noise, an improved IMM filter is constructed to further improve the robustness using the maximum correntropy criterion. The simulation results show that the proposed approach can effectively suppress the non‐Gaussian noise and improve the accuracy with adaptability and robustness.