Abstract:

Oceans are not only super-ecosystems but also strategic resource reservoirs, and thus, ocean monitoring is crucial. Unmanned surface vehicles (USVs) are new types of multifunctional unmanned platforms for ocean monitoring, and path planning plays a crucial role as a core technology in their operation. With the continuous increase in maritime traffic density and upgrading of navigation safety standards, traditional path planning methods are facing growing challenges in adapting to complex environments. In this study, a multidimensional improvement strategy is proposed to address the limitations of the bidirectional rapidly-exploring random tree star(Bi-RRT^*) algorithm in USV path planning. First, an adaptive step-size adjustment mechanism, based on environmental feature perception, is established; second, a key node selection strategy is designed; and finally, Bezier curves are used to smooth the generated path, producing a smoother trajectory that better meets the kinematic requirements of USVs. Simulation results show that the improved bidirectional RRT^* algorithm outperforms its traditional counterpart in terms of node-generation efficiency, overall performance, and path smoothness.

Key words: unmanned surface vehicles, ocean monitoring, path planning, bidirectional RRT^*, adaptive step-size, Bezier Curve, key node selection