The PMHT for maneuvering targets

Abstract

Via the EM algorithm and a slight modification of the usual target-tracking assumptions, the probabilistic multi-hypothesis tracker (PMHT) of Streit and Luginbuhl (1995) combines data-association and filtering to a simple, elegant, and efficient iterative procedure. The PMHT works well; but part of its appeal is a consistent and extensible statistical foundation. In the paper we show one such extension, that designed to track maneuvering targets. The basis, as is in common with many algorithms designed for maneuvering targets, is of an underlying and hidden "model-switch" process controlled by a Markov probability structure.