Visual Quality Measures for Characterizing Planar Robot Grasps

Abstract

Abstract-This paper presents and analyses twelve quality measures that characterize robotic grips according to their stability and reliability. The measures are designed to assess three-finger grips of 2D parts performed in a real environment, taking into account both theoretical aspects and unavoidable uncertainties of a grasping action. They build on the existing literature and on physical and mechanical considerations. The measures constitute a feature space that pattern recognition methods can use in order to classify robotic grips according to their quality. Six of the measures depend on the actual finger configuration of the gripper, and they have shown to be critical for a better characterization. The kinematics of the Barrett Hand has been used. As a validation step, the measures are merged in two global quality values (with different practical applicability) that can be used to rank feasible candidate grips.

Biologically-inspired 3D grasp synthesis based on visual exploration

“…The grasp is selected according to visual criteria [11] on the regions of the object that have received more attention during the reconstruction process. Figure 11 shows some examples of grasps executed on these regions.…”

“…Alternatively, some authors use heuristics to reduce the number of candidates during the grasp synthesis and obtain a good grasp in short time [4]; others approximate an object model with a set of shape primitives -such as cylinders, boxes, or cones-and use rules, based on those primitives, to generate grasp pre-shapes and starting positions [26]. Quality measures for grasp analysis have been described in several works [5,11].…”

“…If the available information is enough to perform the above evaluation, we use a number of heuristic criteria that we have previously defined and validated experimentally [11] to decide if a reliable grip can be extracted from it. Some of the criteria are related to the hand kinematics, but in this case we want to keep the process purely visual, and as flexible as possible.…”

“…The system, in its aim of generating possible grips, assesses the quality of the available data through the use of visual criteria [11]; based on these data, it parameterizes the further exploration of areas that are considered interesting for grasping but not reliably covered by the visual analysis. Therefore, the decision on whether to carve a voxel or not in algorithm 2 is related not only to the voxel occupancy, but also to the fulfillment of the grasping criteria.…”

Evaluation of Human Prehension Using Grasp Quality Measures

“…There have been some attempts in robotics to combine some of the measures so as to create global quality indices [4]. This was done by using the sum (or weighted sum) of a set of selected quality measures in a single global index, considering that all of them have to be either maximized or minimized [5][6][7][8][9]. These works showed the problem of merging indices with different numerical ranges and physical units.…”