King's College London

Research portal

Robotic Untangling of Herbs with Parallel Grippers

Research output: Chapter in Book/Report/Conference proceedingConference paper

Original languageEnglish
Title of host publication2020 UK Robotics and Autonomous Systems Conference (UKRAS)
Subtitle of host publicationRobots into the Real World
PublisherThe Engineering and Physical Sciences Research Council (EPSRC)
DOIs
Publication statusPublished - Apr 2020

King's Authors

Abstract

Robotic packaging generally involves picking target objects from a pile consisting of many other similar or random objects. For a pile composed of herbs, the weight picked up can be controlled by varying the opening aperture width of a parallel gripper. However, the individual strands of herbs get entangled with each other, causing more to be picked up than desired. Here, it is shown that using a spread-and-pick approach the degree of entanglement in a herb pile can be reduced. Compared to the traditional approach of picking from an entanglement-free point in the pile, the proposed approach reduces the variance in picked weight for homogeneous piles of plastic and real herbs by 36.35% and 23.64%, respectively. These results demonstrate that using the proposed spread-and-pick approach, the stochasticity of a herb environment can be counterbalanced effectively.

View graph of relations

© 2018 King's College London | Strand | London WC2R 2LS | England | United Kingdom | Tel +44 (0)20 7836 5454