Gait Control Method of Robot and Biological Systems by Use of Hojo-Brain

Abstract

The purpose of this research is to propose and develop a new control method in the robotic field and the bio-medical field, which is configured by the robotic/biological simulator, analytical motion mode, sensory feedback and the artificial CPG which is constructed by recurrent neural network (RNN) and genetic algorithm (GA). We call such controller “HOJO-Brain”, which means supplementary brain for motion control. We apply this method in the robotic field and the bio-medical field. In the robotic field, this HOJO-Brain is applied to a 5-DOF legged locomotion robot. In the bio-medical field, this HOJO-Brain is applied to animals as the FES (Functional ElectroStimulation) controller. This FES control system with HOJO-Brain would have a possibility to realize more effective and emergent motion control for severely physically handicapped persons such as the quadriplegia. By the computer simulations and the simple actual experiments using animals, we could confirm the fine adaptivity and emergence for the motion control. This HOJO-Brain strategy might be regarded as essential for the total motion control.