Path Planning for Relay Node Selection in Complex Marine Space: A Deep Reinforcement Learning Approach

In complex and dynamic marine environments, traditional autonomous underwater vehicle (AUV) path planning methods face three major challenges: ocean current interference causing trajectory deviation; limited long-distance communication leading to information delays; and insufficient static obstacle avoidance capability. To address these issues, this study proposes an AUV path planning method that considers relay selection under ocean current disturbances. On one hand, a relay cooperation mechanism is introduced to resolve communication and energy bottlenecks in long-distance navigation by optimally selecting relay nodes. On the other hand, the path is dynamically adjusted to adapt to the complex time-varying marine environment, effectively avoiding static obstacles. Simulation results demonstrate that the proposed path planning method can generate a near-global-optimal path under ocean current interference, providing a route that meets navigation requirements with minimal time and energy consumption.