Random Finite Sets and Gaussian Mixture Probability Hypothesis Density Filter in Multi-Target Tracking
Fanbin Meng, Yanling Hao, Weidong Zhou, Feng Jie Sun · 2009
The random finite set (RFS) approach offers a natural and smart means to model multi-target and measurements received by the multi-sensor. The probability hypothesis density (PHD) filter propagates a multi-target statistical first moment, the PHD in place of the full multi-target posterior distribution. But there is no closed form solution to the PHD recursion. The Gaussian mixture probability hypothesis density (GMPHD) filter provides a closed form solution to the PHD filter. The technique is demonstrated to be successful in estimating the correct number of targets and their tracks in high clutter density.