In a large-scale oceanic environment with spatially variable ocean currents, it is vital for unmanned surface vehicles (USVs) to navigate with a safe and energy-efficient path. In this letter, a stream-based VF-RRT* (SVF-RRT*) is proposed for path planning in ocean current fields defined by the stream function. Firstly, SVF-RRT* quantifies the extent to which a feasible path goes against currents with a parameter termed upstream coefficient . Secondly, a heuristic interval is constructed based on the stream function and adjusted by the coefficient. Then, SVF-RRT* employs biased sampling and tree pruning to facilitate convergence. Sampling nodes possessing a stream value outside the interval are rejected with an adaptive probability. Meanwhile, nodes in the exploration tree outside the interval or with a cost higher than the path are discarded. Furthermore, SVF-RRT* utilizes the upstream coefficient to update the extended direction of the tree. Finally, simulation results show that SVF-RRT* accelerates the convergence of upstream cost and generates a smoother path for USVs.
18 August 2025 in Advanced Engineering Informatics
Unsupervised domain adaptation (UDA) aims to enable autonomous vehicles to understand complex road information without semantic labels. However, traditional UDA methods face challenges under adverse conditions, leading to suboptimal segmentation. These methods grapple with two core problems: distinguishing visually similar classes and misclassifying sudden dynamic classes in the target domain. To tackle these issues, we introduce Adaptive Masking and Structure-Constraint Learning (AMSC), comprising: (1) Adaptive Masking module for context leverage. (2) Structure-Constraint Learning to address structural loss from masking. Through these modules, we effectively tackle the aforementioned challenges. Simulation results showcase AMSC's state-of-the-art UDA performance, achieving a mIoU of 73.2 on Cityscapes to ACDC, a standard benchmark for normal-to-adverse semantic segmentation in UDA.