A Dual-LiDAR ship-shore collaborative pose estimation method for USV berthing and unberthing

This paper presents a ship-shore LiDAR collaborative pose estimation method for USVs. The proposed approach enables joint environmental perception by integrating data from shipborne and shore-based LiDAR sensors through point-to-point communication, thereby providing rapid and accurate pose information for the USV. The method comprises three core modules: place recognition, ship-shore relative pose estimation, and factor graph optimisation. Experiments conducted in an open robotic competition simulation environment demonstrate the effectiveness of the proposed system. The ship-shore collaborative pose estimation framework corrects accumulated drift and improves the global consistency of the shipborne trajectory relative to the fixed shore reference. Experimental results indicate that the proposed method attains an absolute trajectory error of 0.42 m and a real-time processing rate of 14.2 frames per second (FPS) in berthing and unberthing scenarios characterised by occlusions and blind zones. These results indicate that the method can support safe and autonomous USV berthing operations in complex port environments, providing critical technical support for the advancement of next-generation intelligent vessels.