Title :
Robotic manipulation of rotating object via twisted thread using high-speed visual sensing and feedback
Author :
Hyuno Kim;Yuji Yamakawa;Taku Senoo;Masatoshi Ishikawa
Author_Institution :
Dept. of Information Physics and Computing, Graduate School of Information Science and Technology, University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, 113-8656, Japan
Abstract :
Manipulation of rotating objects via twisted thread is a challenging task because handling the non-linearity from the flexible object is required. Under the non-linearity, clarifying the physical model of the system and determining its control law become more difficult. In this case, the use of high-speed sensor and actuators can be a solution for the manipulation tasks. Since high-speed sensing and controlling make the system linear in very short time, we can apply more simple model and control law for the tasks. In this paper, we used button spinner as an example and showed the advantages of high-speed visual feedback control for the manipulation. Our contributions are as followings: First, we suggested a visual sensing method to measure the angular position of a rotor at high speed. Second, we achieved the manipulation of the rotating object via twisted thread at high-speed without any sophisticated system model. Finally, we proposed the method to cancel the fluctuation of rotating axis by the high-speed visual feedback. In our method, control of the button spinner was achieved without any force feedback, only with visual feedback and sensing of the joint angle of the robot hand.
Keywords :
"Rotors","Visualization","Robot sensing systems","Image color analysis","Spinning"
Conference_Titel :
Multisensor Fusion and Integration for Intelligent Systems (MFI), 2015 IEEE International Conference on
DOI :
10.1109/MFI.2015.7295819