King's College London

Research portal

Stable Grip Control on Soft Objects With Time Varying Stiffness

Research output: Contribution to journalArticle

Original languageEnglish
Article number7471513
Pages (from-to)626-637
Number of pages12
JournalIEEE TRANSACTIONS ON ROBOTICS
Volume32
Issue number3
DOIs
Publication statusPublished - 18 May 2016

Documents

King's Authors

Activities

Abstract

Humans can hold a live animal like a hamster without overly squeezing despite the fact that its soft body undergoes impedance and size variations due to breathing and wiggling. Though the exact nature of such biological motor controllers is not known, existing literature suggests that they maintain metastable interactions with dynamic objects based on prediction rather than reaction. Most robotic gripper controllers find such tasks very challenging mainly due to hard constraints imposed on stability of closed loop control and inadequate rate of convergence of adaptive controller parameters. This paper presents experimental and numerical simulation results of a control law based on a relaxed stability criterion of reducing the probability of failure to maintain a stable grip on a soft object that undergoes temporal variations in its internal impedance. The proposed controller uses only three parameters to interpret the probability of failure estimated using a history of grip forces to adjust the grip on the dynamic object. Here we demonstrate that the proposed controller can maintain smooth and stable grip tightening and relaxing when the object undergoes random impedance variations, compared to a reactive controller that involves a similar number of controller parameters.

Download statistics

No data available

View graph of relations

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