Tracking Multiple Underwater Targets Using Adaptive Gaussian Mixture Probability Hypothesis Density Filter With Unknown Clutter Rate

This article considers tracking multiple underwater targets using active sonar sensors. The fundamental problem of underwater multiple target tracking is unknown data association between sonar measurements and underwater targets. Gaussian mixture probability hypothesis density (GMPHD) filter yields satisfactory performance for solving the data association problem with its low computational load. When the clutter rate is not available in advance, conventional GMPHD filters have difficulty in tracking targets. To resolve the performance degradation problem of the GMPHD filter in multiple target tracking due to unknown clutter rate, this article addresses an adaptive GMPHD filter. This article assumes that the maximum number of underwater targets in a given workspace is available a priori. Under this assumption, this article presents an adaptive and time-efficient GMPHD, which can handle unknown clutter rate. As far as we know, this article is unique in addressing an adaptive GMPHD for handling unknown clutter rate, such that its tracking accuracy is comparable with the ideal case in which the true clutter rate is available in advance. Moreover, computer simulations show that the time efficiency of using the proposed adaptive GMPHD is comparable with the ideal case where the true clutter rate is accessible.