Multi-AAV Tracking Under Partially Overlapping and Gapped Fields of View in Distributed Multi-Sensor Networks

Le résumé fourni par la source

In distributed multi-sensor multi-target tracking networks with limited fields of view (FoVs), unmanned aerial vehicle (UAV) targets may traverse uncovered regions, causing missing measurements, label inconsistency, and fragmented trajectories. To address this problem, we propose a cross-FoV association-based global track management (GTM) method for consistent and continuous global trajectory recovery. The proposed method classifies fused global tracks into active, pending, and dead sets according to their latest detection times and a gaptolerance threshold, enabling explicit identification of candidate fragmented track pairs. In the association stage, temporal order and predictive Mahalanobis gating are used to link fragmented tracks across FoV gaps while accounting for uncertainty growth during missing-observation intervals. In the reconstruction stage, inconsistent labels are unified and missing states are restored via retrodiction-based interpolation, which exploits both pre-gap and post-gap information for smoother trajectory reconstruction. Simulation results show that the proposed method mitigates FoV-gap-induced degradation and achieves more consistent and continuous tracking performance.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.