A Smooth Global Path Planning Method for Unmanned Surface Vehicles Using a Novel Combination of Rapidly Exploring Random Tree and Bézier Curves

Developing autonomous navigation techniques for surface vehicles remains an important research area, and accurate global path planning is essential. For mobile robots—particularly for Unmanned Surface Vehicles (USVs)—a key challenge is ensuring that sharp turns and sharp breaks are avoided. Therefore, global path planning must not only calculate the shortest path but also provide smoothness. Bézier Curves are one of the main methods used for smoothing paths in the literature. Some studies have focused on turns alone; however, continuous path smoothness across the entire trajectory enhances navigational quality. Contrary to similar studies, we applied Bézier Curves whose control polygon is defined by an RRT path and thus avoided a multi-objective formulation. In the final stage of our approach, we proposed a control point reduction method in order to decrease the time complexity without affecting the feasibility of the path. Our experimental results suggest significant improvements for multiple map sizes, in terms of path smoothness.