Interactive Positioning and Mapping Method for AUV Multibeam Bathymetry

Accurate positioning information serves as the foundation for obtaining high-precision seafloor topographic point clouds using autonomous underwater vehicles (AUVs). Traditional mapping methods directly utilize uncorrected positioning data for beam point calculations, inevitably introducing positioning errors that degrade mapping accuracy. Underwater simultaneous localization and mapping (SLAM) achieves high-precision relative positioning and mapping by leveraging interobservation constraints; however, it partially neglects absolute positioning accuracy. To further enhance the depth measurement accuracy of multibeam mapping for AUVs, this study proposes an interactive positioning and mapping (IPAM) method. Based on real-time monitoring of AUV positioning data quality, a geometric constraint model is constructed with positioning parameters. Overlapping terrain point clouds obtained during stable positioning periods are incorporated as constraints to dynamically solve positioning parameters during unstable periods using a least-squares approach. This iterative correction process ultimately achieves the interactive refinement of both positioning and mapping data. A deep-sea experiment conducted at a depth of 4000 m in the Eurasian Basin of the Arctic Ocean demonstrated that the proposed method effectively reduced point cloud coordinate errors from the decameter level to the meter level. The overall coordinate mean errors of feature points and verification points were reduced by 92.85% and 91.87%, respectively. By introducing an interactive correction mechanism, this method significantly enhances the ability of AUVs to acquire high-precision topographic data in complex deep-sea environments.