Bearing-only tracking of underwater multiple acoustic sources using underwater vehicles under latency and packet loss

Bearing-only tracking (BOT) of underwater acoustic sources is crucial for numerous applications in the underwater domain. A promising approach involves reconfigurable underwater wireless sensor networks (UWSNs) composed of a group of underwater vehicles that enable target state estimation via data fusion while enhancing observability. However, the uncertainty of measurement-target association and unstable acoustic communications pose significant challenges. This article presents a selective association probabilistic multi-hypothesis tracking (SA-PMHT) solution tailored for BOT of underwater targets under latency and packet loss. Specifically, a selective association mechanism is designed to determine measurement-target associations by synthesizing candidate measurements for each target. Then, a pseudo-linear equation is constructed to establish regression matrices for targets, and a sliding-window least-squares estimator is embedded for iterative state estimation. We demonstrate that the SA-PMHT tracking framework is a pseudo-posterior variant of the expectation-maximization (EM) algorithm. Meanwhile, theoretical analyses are provided for its computational complexity and sufficient conditions for target observability. Finally, both simulation and experimental results validate the effectiveness of our solution.

Recommended citation: T. Gao, P. Di Lillo, G. Antonelli, T. Guo, A. Marino, J. Yan, and X. Guan, "Bearing-Only Tracking of Underwater Multiple Acoustic Sources Using Underwater Vehicles Under Latency and Packet Loss," IEEE Transactions on Industrial Electronics, Early Access, 2026.
Download Paper