Obstacle avoidance of multi mobile robots based on behavior decomposition reinforcement learning

摘要

A reinforcement learning method based on behavior decomposition was proposed for obstacle avoidance of multi mobile robots. It decomposed the complicated behaviors into a series of simple sub-behaviors which were learned independently. The learning structures, parameters and reinforcement functions of every behavior are designed. Then, the fusion for learning results of all behaviors was optimized by learning. This learning algorithm could reduce the status space and predigest the design of reinforcement functions so as to improve the learning speed and the veracity of learning results. Finally, this learning method was adopted to realize the self-adaptation action fusion of mobile robots in the task of obstacle avoidance. And its efficiency was validated by simulation results.