Target-Centric Probabilistic Multi-Hypothesis Tracking
Abstract
This report examines an alternative assignment model for Probabilistic Multi-Hypothesis Tracking (PMHT) in which targets are assigned to measurements, in contrast to the original PMHT model in which measurements are assigned to targets. A new target-centric PMHT algorithm is derived and compared to its original measurement-centric counterpart. The relationship between target-centric PMHT and the probabilistic data association (PDA) filter for single-target tracking in clutter was examined, and a PDA-style approximation to the target state covariance matrices for target-centric PMHT is proposed. A target-centric/measurement-centric PMHT hybrid is also proposed to address algorithm performance in the case of closely spaced targets.
Document Details
- Document Type
- Technical Report
- Publication Date
- Jun 12, 2020
- Accession Number
- AD1120646
Entities
People
- Michael J. Walsh
- Tod Luginbuhl
Organizations
- Naval Undersea Warfare Center