Probabilistic algorithms for efficient grasping and fixturing (Q1974460)

Consider an idealized robot hand, consisting of several independently movable force-sensing fingers used to grasp a rigid object with a piecewise smooth boundary with \(n\) points where fingers can be placed. Algorithms are given for selecting a ``strong'' grasp with the least number of fingers using several measures of goodness. Furthermore, an approximation scheme is given that guarantees a grasp quality close to the overall optimal value where fingers are not restricted to preselected points. These problems can be represented by a collection of convex set covering problems where either the cover size is minimized or the scaling factor of an inscribed geometric object is maximized.