Multi-target filtering with linearised complexity
Probability
2017-12-05 v4
Abstract
An algorithm for the estimation of multiple targets from partial and corrupted observations is introduced based on the concept of partially-distinguishable multi-target system. It combines the advantages of engineering solutions like MHT with the rigour of point-process-based methods. It is demonstrated that under intuitive assumptions and approximations, the complexity of the proposed multi-target estimation algorithm can be made linear in the number of tracks and linear in the number of observations, while naturally preserving distinct tracks for detected targets, unlike point-process-based methods.
Cite
@article{arxiv.1404.7408,
title = {Multi-target filtering with linearised complexity},
author = {J. Houssineau and D. E. Clark},
journal= {arXiv preprint arXiv:1404.7408},
year = {2017}
}