Tracking of a move-stop-move target in clutter: A comparison among MM methods

In this paper we compare three different sequential estimation algorithms for tracking a single move-stop-move target in clutter. Bayesian estimation algorithms are taken into account, with a special focus on particle filters (PF). The target can undergo three different motion modes: a stopped target mode, a constant velocity mode and a manoeuvre mode. We analyze a realistic car traffic scenario by considering not only additive Gaussian noise, but also detection probability less than unity and false measurements originated by clutter disturbance. The aim of this paper is to compare the so called PDA-MM-PF (probabilistic data association, multiple model, particle filter) and PDA-MM-APF (probabilistic data association, multiple model, auxiliary particle filter) to the well-established Kalman-based PDA-IMM (probabilistic data association, interacting multiple model) tracking algorithm. Tracking filters ignore a priori information about the true clutter spatial density. Advantages and disadvantages of the proposed algorithms are illustrated and discussed through computer simulations.