Multiplatform Bearing-Only Target Tracking Algorithm Based on Maximum Likelihood Maximum Correntropy Kalman Filter

Multiplatform bearing-only target tracking poses a significant challenge in underwater target tracking. Presently, the field often relies on suboptimal nonlinear filters, which unfortunately exhibit numerical instability when confronted with highly nonlinear systems. Hence, this article suggests employing the maximum likelihood Kalman filtering (MLKF) algorithm for multiplatform bearing-only target tracking. This approach first obtains the maximum likelihood estimate (MLE) of the bearings from multiple nonlinear measurements, which are then converted into the target’s position in Cartesian coordinates for the Kalman filter (KF). The KF algorithm is derived based on the assumption of Gaussian noise. However, in engineering practice, measurement noise is often non-Gaussian, which can lead to a decline in the tracking accuracy of the KF algorithm. To address this, the maximum likelihood maximum correntropy KF (MLMCKF) algorithm is proposed, which utilizes the maximum correntropy criterion (MCC) instead of the original minimum mean square error (MMSE) criterion. Simulation results show that the measurement values transformed by the ML method follow a Gaussian distribution. Moreover, under non-Gaussian measurement noise, the proposed MCMLKF algorithm achieves superior tracking accuracy compared to existing non-Gaussian algorithms.