Probability Hypothesis Density-Based Multitarget Tracking for Proximity Sensor Networks

Qiang Le, Lance M. Kaplan

Abstract: An investigation of the feasibility of a mesh network of proximity sensors to track targets is presented. In such a network the sensors report binary detection/nondetection measurements for the targets within proximity. A new probability hypothesis density (PHD) filter and its particle implementation for multiple-target tracking in a proximity sensor network are proposed. The performance and robustness of the new method are evaluated over simulated matching and mismatching cases for the sensor models. The simulations demonstrate the utility of the PHD filter to both track the number of targets and their locations.