Uncorrelated Conversion Based Filtering for Angle-Only Tracking in Modified Spherical Coordinates

For angle-only tracking (AOT), an unscented filter in modified spherical coordinates (MUF) provides a flexible and effective framework. It has the property to decouple the potentially unobservable range from other elements of the state vector and provides a more accurate approximation than the extended Kalman filter (EKF). However, for angle-only tracking, the performance of the MUF is limited by the accuracy of the first two moments and the linear framework. Because it is an approximation to a linear minimum mean squared error (LMMSE) estimator which is only the best when the target state and the measurement are jointly Gaussian. To improve the estimation performance, this paper proposes an uncorrelated conversion based filter in MSC (MUF-UCF), where we combine the recently proposed uncorrelated conversion based filter (UCF) with the MUF. The UCF augments measurement with its uncorrelated conversion (UC) and replaces the original measurement with the augmented measurement in the LMMSE framework. Because the UCF extracts more information from the measurement, it outperforms the original LMMSE estimator. Hence, MUF-UCF can take advantage of the UCF and the MUF at the same time and gain better performance than the MUK. Simulation results show the effectiveness of the proposed estimator.