自己組織化マップを用いた人工生物の評価値推定手法の提案

抄録

We have proposed an automated design method for the robot’s morphology and controller. Our proposed method by using the evolutionary computation has the problem that computational cost is increased. In this paper, we consider an estimated method for the robot’s fitness values with the self-organizing map. As the results of the computational experiments, it was verified that our proposed method is able to classify the type of the robot’s morphology.