Local path re-planning algorithm based on high-order Bézier curves and collision risk analysis model

Abstract In this article, a new strategy has been proposed to plan the local path of unmanned surface vehicles by combining high-order Bézier curves with collision risk analysis models. Continuous high-order Bézier curves are used as splines for replanning paths, ensuring that the path satisfies curvature continuity and velocity continuity. The collision risk analysis model has been introduced, and the single states of ‘contact or non-contact’ in traditional methods have been eliminated, and the feasibility of the path has been quantified as a safety index. In order to improve the quality and speed of path planning, an improved RBMO algorithm is proposed. In the improved RBMO algorithm, individuals execute targeted clustering strategy based on their fitness, performing guidance, hunting, and jumping search tasks respectively. The simulation results show that the algorithm proposed in this paper effectively achieves local path replanning and can generate continuous, smooth, and safe paths. Moreover, the improved RBMO algorithm exhibits significant superiority in this optimization problem.