Multi-hypothesis tracking is a flexible and intuitive approach to tracking multiple nearby objects. However, the original formulation of its data association step is widely thought to scale poorly with the number of tracked objects. We propose enhancements including handling undetected objects and false measurements without inflating the size of the problem, early stopping during solution calculation, and providing for sparse or gated input. These changes collectively improve the computational time and space requirements of data association so that hundreds or thousands of hypotheses over hundreds of objects may be considered in real time. A multi-sensor simulation demonstrates that scaling up the hypothesis count can significantly improve performance in some applications.
@article{arxiv.1905.07466,
title = {Scaling Data Association for Hypothesis-Oriented MHT},
author = {Michael Motro and Joydeep Ghosh},
journal= {arXiv preprint arXiv:1905.07466},
year = {2019}
}