Mixture Maximum Correntropy Unscented Kalman Filter for Angles-Only Target Tracking Problem

This paper addresses an angles-only target tracking (AoT) problem to track a target using only angle measurements in noisy environments. This is a difficult task due to the uncertainties in the target-observer dynamics, and due to the outliers in the angle measurements. Since these uncertainties are modeled as non-Gaussian (a combination of glint and shot noise), a new approach called the mixture maximum correntropy unscented Kalman filter is proposed. This method involves developing an augmented model that combines state prediction and measurement error to handle the uncertainties of tracking problem. The estimation accuracy of the proposed algorithm is evaluated and compared to that of the UKF and maximum correntropy UKF. The simulation results demonstrate that the proposed algorithm performs better in estimating the target's state while dealing with the non-Gaussian uncertainties.