A Speed-Up Method of Calculating Grasp Quality Measure for Optimal Grasp by a Robot Hand

In this paper, when grasping a target object using a robot hand, we propose the fast algorithm how to calculate the appropriate grasping position. Calculation of grasping stability is essential for calculating the most stable grasping position when grasping an object with a robot hand. Several studies are known regarding calculation of grasp quality measure by which calcu-lated using a convex hull generated by six-dimensional wrench vectors. However, when calculating a convex hull composed by six-dimensional wrench vectors, the amount of calculation increases exponentially as the number of dimensions and points increases. In our proposed method, we used parallel processing to speed up the convex hull algorithm used to calculate the grasp points, and change the grasp points used to calculate grasp stability to speed up the calculation of the optimal grasp position. As a result of performing evaluation calculations on a cube with sides of 50 mm, we achieved a reduction in calculation time of approximately 1/30 to 1/50.