Lobster-inspired Finger Surface Design for Grasping with Enhanced Robustness | IEEE Conference Publication | IEEE Xplore

Lobster-inspired Finger Surface Design for Grasping with Enhanced Robustness


Abstract:

This paper presents a lobster-inspired design of a soft finger's contact surface for grasping with enhanced robustness. The lobsters, while living on the seabed with sedi...Show More

Abstract:

This paper presents a lobster-inspired design of a soft finger's contact surface for grasping with enhanced robustness. The lobsters, while living on the seabed with sediments of various sizes, sources, materials, and life forms, exhibit exceptional capabilities in object manipulation under-water using angled-claws with two fingers. By inspecting the geometric features of the lobster tooth, we proposed a series of finger surface designs molded with silicone. We tested the surface friction using traditional methods to shortlist our design pool, and further verified their performance using robotic arm grasping against a series of challenging objects from the EGAD, the Evolved Grasping Analysis Dataset. Results show that, in certain cases, the lobster-inspired finger surface design yields an enhanced grasping success rate by 56% at most than those without the surface. Furthermore, we propose a minimum setup for robotic grasping using NVidia Jetson Xavier, Intel RealSense D435, and the proposed soft gripper to be compatible with most robotic manipulators as a cost-effective configuration for shareable and reproducible research.
Date of Conference: 12-16 April 2021
Date Added to IEEE Xplore: 12 July 2021
ISBN Information:
Conference Location: New Haven, CT, USA

Funding Agency:


I. Introduction

The contact surface design plays a key role to influence the grasping success in object manipulation. The grasp robustness is closely related to the physical contact between the robotic finger and the target object, which is modulated by the amount of force exerted at the finger. Recent research has been devoted to the adoption of model-based [1], [2], [3] or learning-based algorithm [4], [5], [6], [7] development for finding the optimal grasping point or posture, the sensing and control of fingertip while in contact with the object for refined motor function [8], the development of novel grasping mode through mechanism design at the finger [9], as well as the optimal number of fingers for successful grasp [10]. The design and development of an effective finger surface design, even without extra sensing or motor control, remains a challenging research topic that offers a simplified, practical, and cost-effective value in industrial applications.

References

References is not available for this document.