神経振動子を用いた二次元ホッピングロボットの跳躍移動パターン生成

概要

論文の詳細を見る
This study examines the pattern generation system of jumping motion patterns for the two-dimensional hopping robot using the neural oscillators. The hopping robot that has two-link arm makes two-dimensional motion, the hopping height and the moving distance. The proposed pattern generation system composes of the neural network, the central pattern generators (CPGs) and the hopping robot. The CPG is used as the model of the Wilson-Cowan neural oscillator, the parameters of CPGs are determined by the neural network. The weights of the neural network are adjusted by the genetic algorithm (GA) with a mathematical model of the hopping robot. In the experimental system, the internal state of CPG input into each actuators as the torque commands. The effectiveness of the proposed system is demonstrated by experimental results.
香川高等専門学校の論文