Embroidered textile electrodes
: a low-cost interface for electromyography

Student thesis: Doctoral ThesisDoctor of Philosophy

Abstract

Textile electromyography allows for muscle activity sensing through surface electrodes made with fabric, using materials such as conductive inks or threads. These electrodes are promising for healthcare applications. For example, it would allow to control myoelectric assistive devices through clothing without requiring the help of a specialist to place the electrodes. In addition, it would give physiotherapists the opportunity to continuously monitor and assess rehabilitation methods outside of the laboratory.
The embroidered electrodes have been proposed because they are easy to make with a digital sewing machine, low-cost and reusable, making them a potential alternative for populations with limited funds. However, only their functionality has been investigated and there is no evidence of their performance in healthcare applications. In addition, their current method of production could present a limitation in a low-income setting due to the machinery and operation cost.
In this thesis, electromyographic data collected with embroidered electrodes are investigated with the aim of assessing their applicability for myoelectric control. Specifically, the use of embroidered electrodes is investigated for the control of (i) a hand prosthesis for amputees using data collected from muscles of the residual limb and (ii) a knee-ankle-foot-orthosis for polio survivors using data collected from muscles of the healthy contralateral leg. In these, supervised learning methods are used with data collected from embroidered electrodes to classify hand gestures and gait events. Finally, the fabrication of embroidered electrodes by hand is explored in order to propose a solution for populations that cannot afford automated production systems. In this, the electrical properties as well as the functionality of hand-sewn electrodes are analysed.
This thesis shows that embroidered electrodes have performances in human motion classification similar to commercial disposable gel electrodes. Moreover, the hand-sewn electrodes appear to be a viable solution showing performances nearly similar to that of the machine-sewn ones.
Date of Award1 Oct 2021
Original languageEnglish
Awarding Institution
  • King's College London
SupervisorMatthew Howard (Supervisor)

Cite this

'