Passive Maneuvering Target Tracking using 3D Constant-Turn Model

Abstract

Most previous literature in the maneuvering target tracking community focus on 2D models or 3D models where target maneuvers mostly in a horizontal plane, and usually an acceptable accuracy can be expected when these models are applied to 3D tracking of civilian aircraft. However, it is not the case when come to agile military aircraft which turns in the 3D space rather than just horizontally, so more accurate model has to be developed to obtain better estimation of the target's state and that is the emphasis upon which our work places. The main contribution in the paper is two twofold, namely the derivation of the details for 3D constant-turn maneuvering model and its application in passive target tracking. Due to the nonlinear measurement and large initial errors in our application background, classical extended Kalman filter (EKF) degrades its performance greatly and is prone to diverge, for this reason we developed the unscented Kalman filter (UKF) based IMM algorithm to improve the tracking precision. Simulation studies are performed on two illustrative examples and good tracking performance has been achieved with only single passive sensor, which indicates the proposed method is effective.