画像認識と位置・力情報を用いた模倣学習  [in Japanese] Imitation Learning Using Image Recognition and Position / Force Information  [in Japanese]

Access this Article

Author(s)

Abstract

<p>This paper proposes imitation learning using position, force, and visual information. Imitation learning enables robots to substitute for human works. In conventional research, the performance of robots is improved using position and force information in imitation learning. However, imitation learning using all of position, force, and visual information are not verified. Therefore, this research demonstrates imitation learning with visual information in addition to position and force information. By using visual information, tasks that robots can execute are expanded. In this paper, it is verified that the real time behavior decision by deep learning using position, force, and visual information is effective. As a result, autonomous operation of an eraser at unknown height and position was successful by the proposed method.</p>

Journal

  • The Proceedings of JSME annual Conference on Robotics and Mechatronics (Robomec)

    The Proceedings of JSME annual Conference on Robotics and Mechatronics (Robomec) 2019(0), 2A2-C02, 2019

    The Japan Society of Mechanical Engineers

Codes

Page Top