Application of Adaptive Weighted Strong Tracking Unscented Kalman Filter in Non-Cooperative Maneuvering Target Tracking

Abstract

An adaptive weighted strong tracking unscented Kalman filter is proposed in this paper for long-range relative navigation alongside non-cooperative maneuvering targets. First, an equation for obtaining the relative motion of two bodies is derived, it can be well adapted for a problem that has medium or long-distance. Secondly, a variance statistics function is introduced in the method to calculate residual weight in real time. The residual weight can be used to adjust the contribution of different measurement information to the fading factor. In this way, the sensitivity of the system to small pulse maneuvers is improved. Finally, the mean and covariance of the posterior state are calculated by the unscented transformation. A replacement equation for the fading factor is derived to improve the first-order approximation accuracy for a strong tracking system. Impulsive maneuvers with three different magnitudes are employed in a series of tests. Results from different methods showed that the proposed method could effectively detect pulse maneuvers with low latency. The proposed method is also numerically stable.