A feedback approach to multitarget multisensor tracking with application to bearing-only tracking

A novel approach to multitarget multisensor tracking, exploiting the feedback connection of a PHD (Probability Hypothesis Density) smoother and multisensor multidimensional data association, is presented. The PHD smoother is used to initialize target tracks while the feedback from the hard data association makes the PHD smoother, and hence the overall tracker, less sensitive to missed detections and false alarms. An application to bearing-only tracking is investigated in order to demonstrate the potentials of the proposed approach for tracking problems wherein the state of a target is observable from multiple sensors but not from a single one.