An Improved Path Planning Algorithm for USV Based on Voxel Extension

In this paper, we propose a corridor generation algorithm based on voxel extension to obtain a safe path for unmanned surface vehicle (USV). Firstly, a dual-drive model is applied to model the USV, and then we design a path planning framework for the front-end and back-end parts. The front-end structure is responsible for finding a series of asymptotically optimal path points and generating safe corridors connected by a series of convex polygons based on these path points. The back-end part takes the safety corridor as a partial constraint and applies numerical optimization methods to generate a smooth trajectory. Finally, we validate and compare the proposed method with the baseline in a simulation environment.