Page 144 - Control Theory in Biomedical Engineering

CHAPTER 6

Modeling simple and complex

handwriting based on EMG signals

a
b
Ines Chihi , Ernest N. Kamavuako , Mohamed Benrejeb c
a
Laboratory of Energy Applications and Renewable Energy Efficiency (LAPER), Tunis El Manar University,
Tunis, Tunisia
b
Department of Informatics, Center for Robotics Research, King’s College London, London, United Kingdom
c
Laboratory of Research in Automation (LA.R.A), National School of Engineers of Tunis, Tunis El Manar
University, Tunis, Tunisia
1 Introduction
Losing the ability to fully control an upper limb, either because of amputa-
tion or paralysis, has a profound effect on the activities of daily life. In the last
decades, assistive robots and prostheses have been considered important
research areas aiming to allow people with disabilities to regain functions
(Mastinu et al., 2018; Adewuyi et al., 2017). Advances in mechanical tech-
nology over the past two decades have resulted in the development of
advanced robotic systems. However, due to the anatomical and physiolog-
ical complexity of the human hand, fully replicating all its functions, in a
natural and autonomous way, has proven to be highly challenging. One
of the most promising control approaches in prostheses and assistive devices
is based on the electrical activities of either remnant or partially active mus-
cles, called electromyography (EMG) signals (Kuzborskij et al., 2012;
Hincapie and Kirsch, 2009). This has led to hundreds of research studies
focusing on improving grasping and reaching movements (Artemiadis and
Kyriakopoulos, 2010; Scheme et al., 2010; Kamavuako et al., 2013) using
different approaches such as regression and pattern recognition (Chen
et al., 2013; Alimi and Plamondon, 1993; Fischer and Plamondon, 2017).
The movement of the hand is produced by a complex combination of
muscles from the forearm and some intrinsic muscles of the hand itself. Mus-
cle fibers receive a nerve impulse that is converted into energy in the form of
contraction by a chemical reaction at the level of myofibrils. Thus the nerve
impulse results in electrical impulses whose discharge frequency depends on
the required movement (Rouvie `re et al., 1968). EMG is the recording of
muscle electrical potentials providing a window through which motor con-
trol can be investigated. EMG signals are nonstationary, stochastic in nature

Control Theory in Biomedical Engineering © 2020 Elsevier Inc.
https://doi.org/10.1016/B978-0-12-821350-6.00006-8 All rights reserved. 129