Gait Pattern generation of the Quadruped Robot based on composition and decomposition of Motion data

Abstract

In this presentation, the gait patterns of a quadruped robot are analyzed by decomposing and integrating motion by feature quantity analysis based on machine learning from a motion data. Recently, a complication of robot control becomes issues, and the development of a system which can realize various movements with a simple system is required. Despite, there exist a lot of studies are attempting to generate new motion using the quadruped locomotion robot whose controller consists of the central pattern generators, the discussion of how the gait patterns are generated is still open. Here, we consider the possibility of generating new gait patterns by a behavioral approach, in which the behavior data are brought from all of sensors and actuators information online.