This paper proposes a labeled multi-Bernoulli (LMB) tracking method based on generalized covariance intersection (GCI) fusion for multi-target tracking in distributed passive sonar network with sensing coverage constraints. By partitioning sensing coverage sectors between adjacent nodes, the multi-target posterior density is analytically decomposed into overlapping and non-overlapping regions, and a closed-form expression for the fused posterior density is derived. An label matching mechanism leveraging GCI divergence is designed, eliminating the requirement for prior knowledge of nodes’ coverage ranges while effectively addressing label inconsistency and missing associations between adjacent nodes. Simulation results demonstrate that the proposed method achieves robust fusion of multi-node target information under different sensing coverage scenarios, with significant improvements in tracking accuracy and trajectory continuity. SwellEx-96 sea trial data further validates that constraining the effective angular coverage of sonar nodes mitigates performance degradation caused by degraded bearing estimation accuracy in end-fire directions.
Ce et al. (Sun,) studied this question.