油圧モータ・負荷系へのニューラルネットの適用 : 第1報,不感帯動作信号領域の存在する油圧モータ・負荷系の角度制御

概要

論文の詳細を見る
This paper deals with an application of the Neural Network for an angular position control of a hydraulic motor/load system with an undesirable dead zone in small actuating signals. At first, the mathematical model of the hydraulic motor/load system is investigated by comparing the experimental step response with the simulations, and consequently it is confirmed that an initial driving torque with static friction causes the expansion of dead zone. Then, the angular position control system following the reference model response is constructed by using the Neural Network compensator which adjusts the state feedback gains accoding to the variations of the control conditions. By carrying out the experiments and simulations to evaluate the control accuracy and the robustness, the significant effects of the Neural Network compensator for improving the control characteristics are recognized. The only exception would be the small command angles that correspond to the small actuating signals near to the dead zone.
一般社団法人日本機械学会の論文
2001-07-25