Object transportation by two humanoid robots using cooperative learning

In this paper, we propose an approach to the behavior acquisition required for humanoid robots to learn a cooperative transportation task. In case of object transportation with two humanoid robots, mutual position shifts may occur due to the body swinging of robots. Therefore, it is necessary to correct the position in a real-time manner. Many efforts are needed to develop the position shift correction system. We propose to solve the problem by learning required behaviors with two learning algorithms. Successful cooperation of two HOAP-1 humanoid robots in the transportation task obtained by classifier system and Q-learning has been confirmed experimentally in our work.