Research on AGV Path Planning Based on Improved Q-learning
For addressing the slow convergence and dynamic path adjustment issues in AGV path planning,an enhanced Q-learn-ing algorithm is proposed.Firstly,the Manhattan distance is introduced as additional heuristic information,combined with the Q-learning algorithm for path planning to accelerate the convergence speed of the algorithm.Secondly,the consideration of fault points is added,and the path is dynamically adjusted during the path planning process,validating the algorithm's feasibility for dynamic environments.Additionally,a mechanism for collecting goods along the path is designed,allowing the AGV to perform cargo transportation tasks while executing its main tasks.Through comparative experiments,the effectiveness and performance ad-vantages of the improved algorithm in various scenarios are verified.The experimental results demonstrate significant improvements in convergence speed,adaptation to complex environments,and flexible task execution,providing a novel solution for AGV path planning.
万方数据
维普
