Unmanned Ship Path Planning Based on Improved Genetic Algorithm

In order to solve the problems of excessive steering angle and high collision risk when intelligent ships plan paths in restricted waters such as inland waterways, port space and strait channels, this paper proposes a path planning algorithm based on improved genetic algorithm. This method designs a restricted random initialization method to generate an initial population to generate multiple collision-free paths, improves the quality of the initial population, and designs adaptive crossover and mutation operators to improve the convergence ability of the algorithm. For multi-objective optimization problems, the Pareto-binary tournament selection operator is used. According to the shortest path, the best smoothness and the safest path, half of the population is selected for crossover and mutation. The security operator is used to keep the path outside the safe distance. Taking the shortest path and the optimal smoothness as the fitness function, the unmanned ship path planning system is established. The deletion operator is used to delete the redundant path points, and the security operator is further used to regenerate the path points that are not kept outside the safe distance.