The proposed work focuses on the path planning for Unmanned Surface Vehicles (USVs) in the ocean enviroment, taking into account various spatiotemporal factors such as ocean currents and other energy consumption factors. The paper proposes the use of Gaussian Process Motion Planning (GPMP2), a Bayesian optimization method that has shown promising results in continuous and nonlinear path planning algorithms. The proposed work improves GPMP2 by incorporating a new spatiotemporal factor for tracking and predicting ocean currents using a spatiotemporal Bayesian inference. The algorithm is applied to the USV path planning and is shown to optimize for smoothness, obstacle avoidance, and ocean currents in a challenging environment. The work is relevant for practical applications in ocean scenarios where an optimal path planning for USVs is essential for minimizing costs and optimizing performance.

翻译：本文聚焦于海洋环境中无人水面艇（USV）的路径规划问题，综合考虑了洋流等多种时空因素及其他能耗影响因子。研究提出采用高斯过程运动规划（GPMP2）——一种在连续非线性路径规划算法中展现出良好效果的贝叶斯优化方法。通过引入基于时空贝叶斯推理的新因子，实现对洋流的跟踪与预测，从而改进了GPMP2算法。该算法被应用于USV路径规划，在复杂环境中实现了对路径平滑性、障碍规避及洋流优化的兼顾。本研究对实际海洋场景中需通过最优路径规划以最小化成本并优化性能的应用具有重要参考价值。