手動制御系における人間オペレータの前後におけるゲインと人間の学習特性について

概要

論文の詳細を見る
Considering human being as a control element in a feed-back control system, we use as input random noise, in the form of electrical voltage, generated by a noise-generator. We use as controled system 1, 1/s 1/s+1, 1/s (s+1), 1/s^2, using an analogue computer, and a potentiometer having circularly rotating handle at operation point. We perform experiments in error-compensating manual control system. We let an amateur perform error-compensating operations until he is considered to be sufficiently proficient, and we analyze its process. What is unique in this study is that we consider the gain which is obtained by taking the difference between the circular transfer function including the human operator and human transfer function; and that we study the human characteristics when we change the gain before and after the Human operator, with the product of the two gains kept constant. We obtained better control results for controled systems of first order than those of zero order. We found also that for controled system of second order it is likely to diverge and it requires considerable proficience to control. We think it is impossible to controled systems of more than second order. For error e，we take control area ∫^T_O |e|dt and ∫^T_O e^2dt，and we state progress of learning as effect of human learning.
琉球大学理工学部の論文
1972-03-00