Recurrent neural network with self-adaptive GAs for biped locomotion robot

We propose a method for generating stable motion of a biped locomotion robot. We apply the proposed method to eight force sensors at the soles of the biped locomotion robot. The zero moment point (ZMP) is well known as the index of stability in walking robots. ZMP is determined by the configuration of the robots. When we use ZMP as the stabilization index, we must select the best among many stability configurations. Then it is a problem of which configuration is selected. In this paper, the problem is solved with a recurrent neural network. We calculate the position of ZMP and the joints and the angles that should be actuated can be determined by the recurrent neural network without ZMP moving out from the supporting area of the sole. We employ a recurrent neural network with self-adaptive GAs for its learning capability. Further, we built a trial biped locomotion robot, which has 13 joints and verified that the calculated stability motion trajectory can be successfully applied to practical biped locomotion. In this paper, we propose a way of training the recurrent neural network for a biped locomotion robot