Analytic grasp success prediction with tactile feedback

Abstract

Predicting grasp success is useful for avoiding failures in many robotic applications. Based on reasoning in wrench space, we address the question of how well analytic grasp success prediction works if tactile feedback is incorporated. Tactile information can alleviate contact placement uncertainties and facilitates contact modeling. We introduce a wrench-based classifier and evaluate it on a large set of real grasps. The key finding of this work is that exploiting tactile information allows wrench-based reasoning to perform on a level with existing methods based on learning or simulation. Different from these methods, the suggested approach has no need for training data, requires little modeling effort and is computationally efficient. Furthermore, our method affords task generalization by considering the capabilities of the grasping device and expected disturbance forces/moments in a physically meaningful way.

Grasp quality evaluation done right: How assumed contact force bounds affect Wrench-based quality metrics

“…As discussed in [15], this entails approximating the contact surfaces C i by projecting the corresponding taxel grids from the curved finger pads onto planes tangential to the pressureweighted centers p i . Then, the exertable contact forces and moments (which differ between contacts in general) are obtained and used to compute the contact wrench set in (2).…”

“…Computation of q * in [24] requires, for each wrench in (8), the solution of a determinant maximization problem with linear matrix inequality constraints. In [15] we proposed an efficient computation of q * for independently bounded grasp contact forces as the solution of a Linear Program (LP) for which highly efficient solvers exist. The formulation expresses each of the scaled task wrenches in (8) as convex combinations of the discrete grasp contact wrenches in (3) summed over all contacts q * L∞ ∈ arg max q L∞ ∈R+, λi,j∈R…”

“…The lifting task was modeled with a wrench set in (8) corresponding to the gravity force distributed in a radius of 6.7 mm around the object's center of mass to account for perception uncertainties (see [15] for a sensitivity analysis regarding the task parametrization). For a more detailed description of the experimental setting and the used parameters we refer the reader to [15].…”

“…In our experiments, the q * L∞ metric yielded 25 % false positives over all trials opposed to only 2 % for the q * L1 metric. Nevertheless, by simply shifting the decision boundary, success prediction with the q * L∞ metric can be made as conservative as desired (see [15]). …”

“…However, analyzing dynamic stability is difficult in practice as it requires to include local contact curvature, mechanical compliance, hand control laws (a compliant controller which allows hand-relative object movement results in a change of GWS) as well as the magnitude and arrangement of the applied contact forces [13], [14]. Nevertheless, our previous work [15] showed that realistic contact force modeling, involving tactile feedback and joint effort limits, allows wrench-based reasoning to predict grasp success surprisingly well and on-par with recent approaches based on supervised learning [9], [16], [17].…”