提出了一种动态环境下多个机器人获取合作行为的强化学习方法,该方法采用基于瞬时奖励的Q-学习完成单个机器人的学习,并利用人工势场法的思想确定不同机器人的学习顺序,在此基础上采用交替学习来完成多机器人的学习过程。试验结果表明所提方法的可行性和有效性。 This paper proposes a reinforcement learning method for multiple robots to acquire cooperative behavior under dynamic environment. The method uses Q-learning based on instantaneous reward to complete the learning of single robot and uses the idea of ​​artificial potential field method to determine the learning sequence of different robots. Based on the use of alternating learning to complete multi-robot learning process. The test results show the feasibility and effectiveness of the proposed method.