Efficient Path Planning for Unmanned Surface Vehicles Using DGDS-RRT* Algorithm

The traditional RRT algorithm and its variants have shortcomings such as slow convergence, unsmooth paths, and inefficient obstacle avoidance. To address these issues, this paper proposes the Dynamic Gradient-guided Dual-strategy RRT* (DGDS-RRT*) algorithm. Its key improvements include: Firstly, a dynamic gradient-guided sampling strategy that balances global exploration and local convergence by adjusting sampling weights and scales. Secondly, a dual expansion strategy combining Goal-biased expansion and improved APF expansion to enhance robustness, and a multi-factor parent node reconnection strategy optimizing path length, smoothness, and safety. Finally, a three-stage path smoothing strategy to generate continuously differentiable trajectories. DGDS-RRT* is used to plan the travel path of unmanned surface vessels. The simulation results show that the performance of DGDS-RRT* is superior to that of BI-RRT*, BI-APF-RRT*, BAI-RRT* and GB-RRT*.