Grasp the Possibilities: Anthropomorphic Grasp Synthesis Based on the Object Dynamic Properties

Abstract

The synthesis of optimal grasp configurations of heavy objects with anthropomorphic hands benefits from the explicit consideration of the object's dynamic properties. A new solution for the fast synthesis of anthropomorphic grasps that ensures the reduction of gravitational and inertial effects during the execution of manipulation tasks, as well as an equal distribution of the grasping forces among all fingers, is proposed in this article. The reduction of the computational complexity is achieved by considering only the regions of the object surface favoring the synthesis of minimal inertia grasps. Moreover, the hand and fingertip size, the hand kinematics, the object model uncertainty, and the surface curvature are all employed to further reduce the number of discrete grasping regions selected for the computation of the optimal grasp configurations with respect to a number of grasp quality indices. The effectiveness of the proposed method has been demonstrated with several case studies.