Limits of linear multitarget tracking

Abstract

Multitarget tracking in clutter has two levels of complexity. One is caused by the exponential increase of number of measurement histories in time, and the other is caused by complexity in allocating measurements to tracks in each scan, which is also exponential in the number of tracks and the number of measurements involved. Linear multitarget tracking is a Bayessian method for multi target tracking which dispenses with measurement to track allocation completely. This results in complexity which is linear in the number of tracks and the number of measurements. Linear multitarget tracking is a recent development, with published results in a limited environment of just two targets in heavy clutter. This paper presents a simulation study which investigates the limits of linear multitarget tracking, both in the number of targets and the clutter measurement density. False track discrimination and target retention statistics are presented and compared with other single target and multi target tracking algorithms.