|
Currently, the incorporation of improved population initialization methods and strategies for adaptively adjusting crossover and mutation probabilities into genetic algorithms for mobile robot path planning has made some progress. However, certain issues still remain, such as poor quality of the initial population and improper setting of genetic algorithm parameters (e.g., crossover and mutation probabilities). These issues lead to slow algorithm convergence and difficulty in balancing multi-objective weights. To address these challenges, this paper proposes a path planning method that integrates the Ant Colony Algorithm with an improved Genetic Algorithm. The method first utilizes the Ant Colony Algorithm to plan multiple feasible paths, which are then used as the initial population for the Genetic Algorithm. Next, the fitness function considers path length, number of turning points, and travel time, and is used as the criterion for evaluating the optimal path. The multiple paths generated by the Ant Colony Algorithm are then ranked according to their fitness values, and a selected number of paths undergo adaptive crossover and mutation. Finally, a redundant point elimination strategy is employed to refine the paths, achieving obstacle avoidance and global path planning for the mobile robot. Simulation tests on a grid map demonstrate that, compared to traditional Genetic Algorithms, improved Genetic Algorithms, and other proposed improved algorithms, the method presented in this paper effectively shortens the path length and reduces the number of convergence iterations, showcasing advantages in both efficiency and stability.
|
