1A1-Q10 操作型作業機械の知能化に関する研究 : 第10報:手先の外力・移動方向を用いた物体把持推定の高精度化(建設・解体ロボット・メカトロニクス) 1A1-Q10 Research on Intelligent Operated-Work Machines : Accuracy Improvement of Grasp-Estimation Using Force and Movement Vectors(Robotics and Mechatronics in Construction and Demolition)
In this paper, a method to accurately estimate whether a grapple grasps an object or not for construction machinery is proposed on the basis of extending the requisite condition for grasp, considering robustness and practicality. Accurate grasp estimation requires confirmation that the hand and arm loads continue to be observed during arbitrary manipulator movements. Enhancement conditions are thus defined by using force and movement at the end-point, including (i) vertical downward force, (ii) longer direction movement, and (iii) horizontal direction reciprocating movement. In transport experiments conducted by using an instrumented hydraulic arm, results indicated that the ratio of error estimation decreases as formed conditions increase, and confirmed that the proposed method estimates grasp/non-grasp accurately and robustly.
ロボティクス・メカトロニクス講演会講演概要集 2013(0), _1A1-Q10_1-_1A1-Q10_4, 2013