Interacting Multiple Models Combined with Probabilistic Data Association for Tracking in Clutter with Particle Filtering

Abstract

This paper describes a novel wayof introducing Interacting Multiple Model -IMM - and Probabilistic Data Association -PDA - into tracking algorithms using ParticleFilters - PF. The benefits of IMM and PDAhave been proven extensively in classic Kalmanfilters and are explained here. Despite the factthat this two tracking improving methods havebeen applied to PF before, we propose an algorithmwhich takes advantage of the optimalimportance function.