A Markerless Human-Manipulators Interface Using Multi-Sensors.
Ping Zhang,Xin Liu,Guanglong Du,Bin Liang,Xueqian Wang
DOI: https://doi.org/10.1108/ir-03-2015-0057
2015-01-01
Industrial Robot the international journal of robotics research and application
Abstract:Purpose – The purpose of this paper is to present a markerless human–manipulators interface which maps the position and orientation of human end-effector (EE, the center of the palm) to those of robot EE so that the robot could copy the movement of the operator hand. Design/methodology/approach – The tracking system of this human–manipulators interface comprises five Leap Motions (LMs) which not only makes up the narrow workspace drawback of one LM but also provides redundancies to improve the data precision. However, because of the native noises and tracking errors of the LMs, the measurement errors increase over time. To address this problem, two filter tools are integrated to obtain the relatively accurate estimation of the human EE, that is, Particle Filter for position estimation and Kalman Filter for orientation estimation. Because the operator has inherent perceptive limitations, the motions of the manipulator may be out of sync with the hand motions, so that it is hard to complete with the high performance manipulation. Therefore, in this paper, an over-damping method is adopted to improve reliability and accuracy. Findings – A series of human–manipulators interaction experiments were carried out to verify the proposed system. Compared with the markerless and contactless methods(Kofman et al., 2007; Du and Zhang, 2015), the method described in this study is more accurate and efficient. Originality/value – The proposed method would not hinder most natural human limb motion and allows the operator to concentrate on his/her own task, making it perform high-precision manipulations efficiently.