Multi-robot Cooperation Strategy Based On Behavior And Learning Mechanisms

Multi-robot systems have broad application prospects in industry, agriculture, defense and other fields. Multi-robot system can deal with tasks that are difficult to be accomplished by an individual robot and according to cooperation, a team of robots may perform the task in a more reliable, faster, or cheaper way. Therefore, the research on coordination and cooperation of multi-robot systems has significant theory and application value.The architectures of multi-robot systems and robot individual are studied in this paper. A hierarchical architecture is presented for multi-robot systems based on the research and analysis of the distributed architecture of multi-robot systems; for the robot individual, a reactive architecture is designed.A coordination collision avoidance method for multi-robot systems in unknown environments is studied firstly. The basic behavior collection of robot individual is presented based on the principle of vector potential field, and the strategy of robots collision avoidance is realized according to the traffic rules; the basic behaviors' control parameters of the reactive architecture are improved and optimized. The method this paper presented not only maintains the inherent advantages of the behaviors based on the theory of potential field, but makes the behavior-based coordination in unknown environment more effective.Multi-robot cooperative collection is studied based on behavior learning. When robot move near the target, the behavior of slow movement is applied, which realizes the queue coordination and avoid the conflict and deadlock; Q-learning is designed and improved the by presenting a kind of process reward, which advances the efficiency of multi-robot systems and the fitness of behavior; the share zone of the optimal strategies is presented, which not only enhances the coordination effect, but improves the overall learning effect of the multi-robot systems; Q(λ) -learning is presented for the behavior integration mechanism which is eased the local shocks ; this paper gives the simulation of all the algorithms and the strategies in JAVA environment.