ニューロン回路 : 運動制御(<小特集>ニューロコンピューティング)

概要

論文の詳細を見る
Recently neurocomputing has attracted attention as a new powerful method of computation. The features of neurocomputing are considered as high-speed and flexible performance by parallel processing and by the ability of learning and adaptivility. In the present article, we briefly introduce applications of neurocomputing to motor control and learning. Neurophysiologically the functions of motor control system, even neuronal circuitry of cerebellum, have been clarified by many researchers. Here we overview these physiological findings and review some attempts of applying computational methods of neural network, such as CMAC, the error-back propagation method and the cerebellar neural network model, to control and learning of robotic manipulators.
日本シミュレーション学会の論文
1989-03-15