Multisensor Swarm Tracking Using Labeled Multi-Bernoulli Filters: A Swarm Geometry-Informed Approach
IEEE Transactions on Aerospace and Electronic Systems · 2026 · 62 · 19409-19421
DOI: 10.1109/taes.2026.3741814 ↗Kurz gefasst
Fachartikel in der Zeitschrift IEEE Transactions on Aerospace and Electronic Systems, 2026. Verfasst von Tharani Rajapaksha und 3 weiteren Autor:innen. Die Arbeit nennt 34 Literaturangaben.
Abstract (Originalsprache)
Tracking individual members of a dense swarm poses significant challenges due to measurement ambiguity, particularly with sensors that have a limited field of view and resolution. To address this challenge, this article first introduces a sequential Monte Carlo swarm-tracking filter that incorporates swarm geometry information into filtering for accurate tracking of the swarm as an extended object. In this formulation, the likelihood is obtained by marginalizing the single-target detection over the swarm region via Monte Carlo sampling. Then, we propose a centralized multisensor fusion framework tailored for swarm tracking, in which the swarm tracker and the swarm-information-incorporated labeled multi-Bernoulli (LMB) filter operate jointly to estimate both individual target states and swarm information in terms of dynamic and formation-related parameters. In this framework, locally updated posterior densities from the swarm tracker are fused using generalized covariance intersection with adaptive sensor weights, and the resulting fused posterior is incorporated into the local single-target likelihood function of the LMB update to improve individual target-state estimation within the swarm. To evaluate performance, numerical experiments are conducted using bearing-only sensor measurements with range-dependent probabilities of detection and false alarms. The effectiveness of the proposed approach for individual target-state estimation is assessed against the state of the art. The results indicate that the proposed approach returns tracking results that are more accurate in terms of cardinality and state estimates and trajectory continuity.
Zitieren
APA
Rajapaksha, T., Gostar, A. K., Blair, A., & Hoseinnezhad, R. (2026). Multisensor Swarm Tracking Using Labeled Multi-Bernoulli Filters: A Swarm Geometry-Informed Approach. IEEE Transactions on Aerospace and Electronic Systems, 62, 19409-19421. https://doi.org/10.1109/taes.2026.3741814
Vancouver
Rajapaksha T, Gostar AK, Blair A, Hoseinnezhad R. Multisensor Swarm Tracking Using Labeled Multi-Bernoulli Filters: A Swarm Geometry-Informed Approach. IEEE Transactions on Aerospace and Electronic Systems. 2026;62:19409-19421. doi:10.1109/taes.2026.3741814
BibTeX
@article{Rajapaksha2026,
author = {Rajapaksha, Tharani and Gostar, Amirali Khodadadian and Blair, Aidan and Hoseinnezhad, Reza},
title = {Multisensor Swarm Tracking Using Labeled Multi-Bernoulli Filters: A Swarm Geometry-Informed Approach},
journal = {IEEE Transactions on Aerospace and Electronic Systems},
year = {2026},
volume = {62},
pages = {19409--19421},
doi = {10.1109/taes.2026.3741814}
}
EUVIMED