Nonhomogeneous Clutter Environment-oriented Adaptive Probability Hypothesis Density Filter

Abstract

The Unmanned Aerial Vehicle (UAV) plays more and more important role in modern war. It is becoming a new efficient mode to use unmanned reconnaissance aircraft to obtain accurate coordinate of enemy ground target and to provide it for attacked weapon system timely. In the traditional target tracking, it is usually assumed that the clutter is homogeneously distributed in the surveillance region. Actually, the clutter obeys nonhomogeneous distribution in many situations, such as in the ground target surveillance environment, and the performance of the standard probability hypothesis density (PHD) filter would reduce sharply in this situation. A nonhomogeneous clutter environment-oriented adaptive probability hypothesis density (APHD) filter is proposed to solve the problem. The Affinity Propagation (AP) clustering algorithm and the algorithm of finding the convex hull are utilized to confirm the clutter region adaptively. It can achieve the speediness and the high precision of the algorithm by determining the clutter region and measurements adaptively. The simulation results indicate that, in the nonhomogeneous clutter environment, the performance of the proposed method is better than the standard PHD filter.