Evaluation of fast MHT algorithms

Abstract

Over the last few years there has been considerable interest in developing highly effective target tracking algorithms. This interest in robust tracking techniques is spawned by the need to process data that often includes a large number of false detections and missed detections. These conditions are amplified by the need to detect and track increasingly smaller targets as well as targets in harsh clutter conditions. Such ambiguities serve to further exacerbate the tracking problem which must already produce plot to track assignments for closely spaced and crossing targets. With the availability of increased processing power, techniques such as multiple hypotheses tracking (MHT) that were traditionally viewed as impractical, due to the large computational requirements, are now receiving considerable attention due to their ability to handle the ambiguities. This paper discusses the implementation of a fast (MHT) algorithm for robust tracking of radar targets in the presence of the false detections, missed detections, and crossing targets.