ヒューマンダイナミックスの特性写像とその制御性の評価(第2報) : 立位姿勢制御系のアトラクタと神経回路網による特徴抽出

概要

論文の詳細を見る
If the state of a neural network can be ascertained at a certain time, then the state of the human behavior can be evaluated by comparing this with an average state. In the present research, based on the measurement of swaying condition in the standing position, we concentrate on a quantitative evaluation of a human stance posture control system and consider the attractor of a center of gravity, and thereb propose a new method for extracting its system characteristics. The posture control system is identified by using a neural network with an error back propagation method. Outputs from the network can estimate the difference of subjects and also the characteristic posture behaviors at different time. The proposed approach well indicates the human stance condition.
公益社団法人精密工学会の論文
2000-03-05