Evaluation of different robotic grippers for simultaneous multi-object grasping

Abstract

For certain tasks in logistics, especially bin picking and packing, humans resort to a strategy of grasping multiple objects simultaneously, thus reducing picking and transport time. In contrast, robotic systems mainly grasp only one object per picking action, which leads to inefficiencies that could be solved with a smarter gripping hardware and strategies. Development of new manipulators, robotic hands, hybrid or specialized grippers, can already consider such challenges for multi-object grasping in the design stages. This paper introduces different hardware solutions and tests possible grasp strategies for the simultaneous grasping of multiple objects (SGMO). The four hardware solutions presented here are: an under-actuated Constriction Gripper, Linear Scoop Gripper suitable for deform-able object grasping, Hybrid Compliant Gripper equipped with mini vacuum gripper on each fingertip, and a Two-finger Palm Hand with fingers optimized by simulation in pybullet for maximum in-hand manipulation workspace. Most of these hardware solutions are based on the DLR CLASH end-effector and have variable stiffness actuation, high impact robustness, small contact forces, and low-cost design. For the comparison of the capability to simultaneously grasp multiple objects and the capability to grasp a single delicate object in a cluttered environment, the manipulators are tested with four different objects in an extra designed benchmark. The results serve as guideline for future commercial applications of these strategies.

Keywords: hand design; multi-object grasping; robotic grasping; robotic hand; variable impedance.

FIGURE 1

Multi-object grasping with different end-effectors. From left to right: Constriction Gripper, Linear Scoop Gripper, Hybrid Compliant Gripper, and Two Finger Palm Hand.

FIGURE 2

Hardware overview of different end-effectors for multi-object grasping (Explanation for rating symbols: = unsatisfactory; - = fair, o = satisfactory; + = well; ++ = excellent); Explanation: “In-hand manipulation” means, that the gripper can move an object in the hand with dexterity to bring it to a different position or orientation. Explanation dimension: Z ist vertical to the tool flange, X is vertical to the main finger motion (TPH: X to adduction direction).

FIGURE 3

Constriction Gripper. Left: First version of gripper. Middle top: CAD model of updated second version of the gripper used in the experiments. Middle down: A simplified comparison of grip force generation between tendon actuated state of the art gripper a) and constriction gripper b). At a) the grip force is the ratio of $r_1$ to $r_2$ multiplied by the $F_s$ , for b) $r_1$ and $r_2$ are similar, so $F_{\mathit{grip}}$ is similar to $F_s$ ; Right: CG with installed nets.

FIGURE 4

Linear Scoop Gripper. Left top: LSG grasping phantom fish. Left bottom: Schematic of parallel gripper vs. LSG mechanism. Right: Description of different components in LSG. Note that the fingers are easy to change, the grey finger is similar to the white one, build only with a different PLA color. In 2 a newer version of LSG with integrated sensors was used for the photo. For all test the version without sensors was used.

FIGURE 5

Hybrid Compliant Gripper. Left: Mold for mini vacuum grippper. Right: HCG with mini gripper mounted on the fingertips.

FIGURE 6

Two Finger Palm Hand (TPH). Left: Grasping two apples between fingers and palm. Middle:TPH with improved net grasping fish phantom. Right: Different components of TPH.

FIGURE 7

Objects used to evaluate single and multi-object grasping. From left to right: baby carrots, shallots, textile spheres representing cherry tomatoes, and wooden cuboids representing packaged goods.

FIGURE 8

Grasping single items. From left to right: Constriction gripper, SoftEnable linear scoop gripper, HCG with suction fingertips, Two finger Palm Hand; Last row shows the space around each gripper to place it in clutter areas.

FIGURE 9

Performance measurements for grippers without direct grasp force measurement. Left: Current consumption of CG depending from the closing velocity for contact measurement; Right: friction torque of LSG. This measurements are needed to control the grasp force of the grippers.

FIGURE 10

Grasping fish phantom: Left: LSG; Middle; HCG with suctioned slope tool; TPH with human lift strategy.

FIGURE 11

Benchmark setup to test single and multi object grasp; right: grasp trials and placing for different grippers. TPH is able to grasp a high number of objects, but the placing should be improved by better opening motion.

FIGURE 12

Multi-object grasping performed by different end-effectors. First (from left, top and bottom): CG grasping 5 wooden cuboids, CG grasping 6 shallots. Second (from left, top and bottom): LSG grasping 3 carrots, LSG grasping 3 shallots. Third (from left, top and bottom): HCG in pre-grasp approach to grasp 3 shallots, HCG after grasping 3 shallots. Fourth (from left, top and bottom): TPH in pre-grasp approach to grasp 4 carrots, TPH after grasping 4 carrots.

FIGURE 13

Mulit-object grasping strategies executed by CG and TPH. Top: CG collects objects in a linear motion for a group and grasp strategy. Bottom: TPH groups objects against the bin’s wall for a environmental constrained grasp strategy.

FIGURE 14

Storage grasp of delicate object. From left to right. The gripper is positioned to grasp the object; Grasping object by adduction grasp; inhand manipulation of object by fingers to place it inside storage; object in storage.