A Multi-Platform Bearing-Only Localization Method Based on Confidence Interval and Error Compression Iteration

Due to the fact that passive underwater acoustic signals can usually only obtain the bearing information of the target and are subject to complex ocean environmental interference, which poses significant obstacles to target localization using a single platform, this study intends to employ multiple unmanned underwater vehicles for cooperative localization of underwater targets. To address the issues of systematic bias in bearing measurements, high nonlinearity, and unknown target motion states in complex underwater acoustic environments, this paper proposes a bearing-only localization method based on confidence interval and error compression iteration. The unknown motion process of the target is mapped into a region expansion process in space, and the cooperative localization of multiple platforms is used to finely compress the regional error, achieving bearing-only localization of underwater targets through gradual region compression.

The core idea of the proposed method is to map the unknown motion state of the target into an expansion process of an uncertainty region. First, bearing-only measurements obtained by multiple platforms at different time instants are used to construct sector regions with systematic bias. Then, a likelihood function is employed to update the probability distribution over the spatial grid, enabling iterative error compression. Meanwhile, the 95% highest posterior density interval is introduced as the statistical output of the localization results.

Simulation experiments were conducted under three different initial scenarios. The results show that, with bounded bearing measurement errors, the proposed method can compress the area of the 95% HPD region from approximately 10⁷ m² to the order of 10⁴ m² within 120 seconds, while the localization error converges from several kilometers to less than 40 meters.