An interacting multipattern data association (IMPDA) tracking algorithm

This paper presents a novel approach for tracking multiple targets using mu ltiple patterns extracted from measurement sequences. The introduction of p atterns leads to a new paradigm of developing high-performance algorithms. A powerful data association algorithm-interacting multipattern data associa tion (IMPDA) is developed. A Markov chain controls the switching behavior o f multipatterns, which allows an interacting operation of multipatterns. Th e IMPDA algorithm also has 'built-in' multiple models which can track both maneuvering and nonmaneuvering targets equally well. (C) 1998 Elsevier Scie nce B.V. All rights reserved.