An Autonomous Recovery Guidance System for USV Based on Optimized Genetic Algorithm

As a kind of flexible and efficient device that can autonomously complete tasks without human intervention, Unmanned Surface Vehicle (USV) has gained increasing attention in the research field recently. Path planning is an essential hotspot in the study of the USV. Unlike traditional robotic path planning, the path planning of the USV needs to consider the dynamic impact of the water environment as well as the constraints of its own vessel’s kinematics. For the sake of enhancing the practical operability of the navigation, an optimized Genetic Algorithm (GA) based on three-dimensional environment modeling is proposed. By simplifying the 3D coordinate, the algorithm can efficiently deal with the avoidance of dynamic and static obstacles. Population initialization is improved to reduce the calculating load, and the Elitism Strategy is combined to ensure convergence. An innovative Sacrifice Strategy and intraspecific hybrid methods are proposed to further increase the genetic diversity and convergence rate. We also propose a new penalty fitness function. Through simulation results and experiments in a water surface environment, the effectiveness and rationality of this method were verified, providing new ideas for path planning research of the USV.