Arctic shipping lanes are gradually opening, creating an urgent demand for unmanned surface vehicles (USVs) capable of safe and efficient navigation in drifting-ice environments. However, dense, highly dynamic sea ice poses significant challenges for existing obstacle-avoidance approaches. This study proposes a dynamically weighted hybrid obstacle avoidance algorithm integrating an improved VO module and an enhanced APF module. The optimized VO method refines the velocity sampling strategy and incorporates DCPA/TCPA-based risk screening to eliminate high-risk candidate velocities. The improved APF method introduces adaptive parameter regulation, virtual-target-based local minimum escape, and historical-velocity-driven oscillation suppression. Furthermore, a real-time dynamic weighting mechanism is designed to balance the contributions of the VO and APF modules according to the instantaneous environmental risk level. Extensive simulation experiments demonstrate that the proposed algorithm achieves reliable collision avoidance performance, high navigation efficiency, and strong environmental adaptability for USVs operating in dynamic Arctic drifting-ice environments.
Bai et al. (Mon,) studied this question.