ニューラルネットワークを用いたロボット用指先力センサの信号処理の研究 : 第1報, 静電容量を検出端に用いた場合

概要

論文の詳細を見る
The robot finger force sensor should be as small as possible in order to install it in the tip of the robot finger. This paper proposes a new robot finger force sensor which is easily fabricated due to its simple structure and uses neural network to process interacting signals and solve nonlinearity of the sensor. For the above purpose, a trial fabrication of the new finger force sensor with capacitance-type clearance detectors was attempted and the experiments for the finger force sensor were carried out. Then it was shown that the sensor worked effectively as a robot finger force sensor. In addition, this new method is more accurate as compared with the conventional method pseud-inverse matrix.
社団法人日本機械学会の論文
1994-07-25