Iterative joint integrated particle filter data association for multitarget tracking in clutter

In tracking closely located multiple targets, the traditional optimal multitarget data association approach such as joint integrated probabilistic data association (JIPDA) faces exponential complexity caused by combinatorial increasing of the number of possible measurement-to-track allocations, which severely limits its applicability. This paper presents an iterative implementation of the Joint Integrated Particle Filter (JIPF) which is particle filter basedl multitarget tracker with the capability of false track discrimination (FTD), and provides the trade off between the performance and computation resources, termed by iterative JIPF (iJIPF). The iJIPF is capable of approximating the single target tracking IPF to multitarget tracking JIPF by traversing the data association tree from the level 0 to full level within finite number of iterations. The assertions are validated by the simulations.