Posture stabilization of robot by visual feedback

Keisuke Fukuoka, Masanori Nagahara, Yutaka Uchimura

研究成果: Article査読


In the future, with progress in robotics, robots will play an important role in the daily lives of human beings. The posture of a robot should be stable so that it does not fall. The posture of robots has been maintained stable mainly by using force and acceleration sensors. However, human beings maintain the stability of their postures on the basis of not only floor reaction force but also visual information. Therefore the use of visual information could be effective in maintaining the stability of the posture of a robot. The rotation invariant phase only correlation (RIPOC) method can be used to measure the rotation angle between two images. The accuracy of the RIPOC method is high in the case of 2D images, for example, in the case of fingerprint recognition and face recognition. However, it is difficult to accurately measure the rotation angle in 3D space by using this method. This paper describes a new extended RIPOC method that can be used to accurately measure the rotation angle in 3D space. By using visual feedback by the proposed method, the posture of robots can be stabilized. Experimental results that confirm the effectiveness of this method are provided in this paper.

ジャーナルIEEJ Transactions on Industry Applications
出版ステータスPublished - 2011 9月 8

ASJC Scopus subject areas

  • 産業および生産工学
  • 電子工学および電気工学


「Posture stabilization of robot by visual feedback」の研究トピックを掘り下げます。これらがまとまってユニークなフィンガープリントを構成します。