Light-Detection-and-Ranging–Based Unmanned Aerial Vehicle Pose Estimation in Ocean Environment with Deep Point Transformer Network

Accurate relative pose estimation between unmanned aerial vehicles (UAVs) and marine vessels is critical for autonomous UAV operations in ocean environments, including ship-based launch and recovery, as well as aerial data collection. Traditional visual-based and Global Positioning System–based systems often suffer from occlusion, lighting variability, and signal degradation over water. This paper proposes a novel approach to estimate the six-degree-of-freedom relative pose of a UAV with respect to a vessel using a single light detection and ranging (LiDAR) sensor. This method leverages the attention-based mechanisms of the point transformer architecture to capture both local and global geometric features from sparse and noisy point clouds acquired in marine conditions. We constructed synthetic and real-world datasets that incorporate ocean-specific noise profiles and validated our model across multiple trajectories and demonstrated robust performance even with partial scans. It is shown that the proposed architecture significantly outperforms traditional point-based baselines in both accuracy and robustness, enabling reliable pose estimation for autonomous maritime UAV operations.