Soft Biomimetic Optical Tactile Sensors for Slip Detection and Grasp Recovery

  • Jasper W James

Student thesis: Doctoral ThesisDoctor of Philosophy (PhD)

Abstract

Tactile sensing gives humans the ability to interact directly with the environment and has resulted in the world being shaped such that it is easy for us to interact with. Tactile sensing is useful in robotic grasping and provides information complementary to vision such as contact forces and surface texture. Bringing tactile sensing to robots is essential to provide human-level touch capabilities. One aspect of human touch is slip detection, which is the ability to detect when something is moving relative to the skin. Slip detection is utilised by humans in a variety of scenarios including determining ideal grasping strength and inferring object properties.

This thesis involves the study of new sensors and methods of slip detection, and an analysis of how slip detection can be used to improve robot touch. Additionally, this thesis explores how one can mimic human sense of touch to improve slip detection in robots. Work is conducted using different variations of the TacTip, a soft biomimetic optical tactile sensor. Experiments are performed with single sensors featuring different skin morphologies and a novel design of tactile robot hand with integrated miniaturised versions of the TacTip. It is demonstrated that the TacTip is effective at real-time slip detection and grasp recovery, and that mimicking structures in the human fingertip elicits incipient slip allowing for early detection of grasp instability.

This thesis seeks to convince the reader that slip detection is as useful a sense to robots as it is to humans and should be included in any future tactile robot system interacting with the cluttered environment around us. Several contributions to the field of tactile robotics are made by furthering knowledge relating to gross slip detection, tactile robot grasping and incipient slip.
Date of Award24 Jun 2021
Original languageEnglish
Awarding Institution
  • University of Bristol
SupervisorNathan F Lepora (Supervisor) & Martin Pearson (Supervisor)

Cite this

'