5U-05
Path Finding of Autonomous Wheel Robot by Reinforcement Learning
○はふぃやんだ らざん,牛田裕斗,佐久間拓人,加藤昇平(名工大)
In recent years, automated warehouse establishment is increasing rapidly, while conventional warehouses have not distinguished yet. This research uses a mobile robot to be applied in conventional warehouse environments as human support by following an employee. The mobile robot has to avoid any obstacles while finding the shortest path to a goal point, in this case, an employee. Currently, we prepared three obstacle patterns to cleared in some conditions. The learning algorithm is Q-learning which fits into a changing environment like a conventional warehouse. The result shows the capability of Q-learning to clear each of the patterns.