A Bootstrapped PMHT with Feature Measurements

Abstract

The probabilistic multi-hypothesis tracker (PMHT), a tracking algorithm of considerable theoretical elegance based on the expectation-maximization algorithm, will be considered for the problem of multiple target tracking with multiple sensors in clutter. In addition to position observations, continuous measurements associated with the unique, constant—and statistically unknown —feature of each target are incorporated to jointly estimate the states and features of the targets for the sake of tracking and classification, leading to a bootstrapped implementation of the PMHT. In addition, the information matrix for the stacked vector of states for all the targets at all the time steps during the observation time is derived.