Cooperation among multiple mobile robots using intention inference

This paper describes a behavior-decision method using intention inference for multiple mobile robots. The robots infer other robots´ intentions from observation of the situation and others´ behavior to settle conflicts and cooperate in accomplishing the tasks. Since this method frees the robots from detailed communication, it saves costs. Also the robots can use communication functions for other purposes. The robots can adapt to dynamic environments because they can change their plans flexibly according to the situation. If the inference rules used by the robots are made similar to the rules humans use, this method can be applied to cooperation between humans and robots. Using the method, an object placement system has been developed. In this system, the robots share tasks, transport an object cooperatively, and avoid collisions. To illustrate the performance of the system, simulations and experiments were carried out