F-1230 視覚と力覚のセンサフュージョンによるロボットのニューロ制御 : 変形物体の操作(G15-2 ロボットの知能化制御)(G15 ロボティクス・メカトロニクス部門一般講演)

概要

論文の詳細を見る
In playing a contact task with a rigid object by a robot, a force/torque sensor is frequently used to detect the contact force/torque. However, only a force sensor is not enough in manipulation of a deformable object by a robot. In the previous study, a sensor fusion algorithm is proposed to control a robot to insert an aluminum peg set up on the robot hand into a hole fixed on a deformable plate. In this paper, an online learning neural network is introduced into the control system, by which the peg can be smoothly and quickly inserted into the deformable hole. To complete the task, a high-speed camera is used and a real time processing algorithm is developed to detect the deformation of the central line of the hole. The effectiveness of proposed method has been demonstrated through experiments.
社団法人日本機械学会の論文
2001-08-22