Page 233 - Control Theory in Biomedical Engineering

214 Control theory in biomedical engineering

complete or useful acceptance of an electrically powered prosthesis. These
patients experimented with a satisfactory combination of comfort, aes-
thetics, and functional parameters. As a result, it was evident that this tech-
nology might be better than body-powered prostheses.
In terms of costs, externally powered prostheses can cost $25,000 to
$75,000 (Resnik et al., 2012), which is 7.5 times more than a body-powered
prosthesis. For this reason, 3D-printed electrical-powered prostheses con-
trolled by electromyography (myoelectrical prostheses), in some cases by
individual voice, or by electroencephalography (EEG) have been developed
to reduce the cost of hardware and prototyping time.

4.4 Myoelectric prostheses

Before the development of externally powered prostheses, devices were
commanded solely by buttons or switches in an open-loop control manner.
However, it was not until the myoelectric prosthesis appeared that this
approach changed. Myoelectric prostheses belong to the electrical-powered
active prostheses spectrum with capability of reading the electromyographic
(EMG) activity from the voluntary movements of muscles (Popov, 1965).
This technology was developed by Reiter in the period 1944–1948
(Weihe, 1998). He used vacuum tubes to create a myoelectric prosthesis
but with huge dimensions. Consequently, his system did not gain clinical
or commercial acceptance. It was until late 1960s that researchers around
the world “re-invented” the technology with the help of the transistor
and its application as an amplifier. The electromyography upper-limb wear-
able device for amputees consisted of several elements to detect the user’s
intention, trying to mimic the natural human movement, from electrodes
for EMG detection, to signal processors and controllers, ending with
actuators for prosthesis movements. Sensors established communication
between human control signals and the upper-limb prostheses. Then, mod-
ern prosthetic hands incorporated surface electrodes located on the skin sur-
face to acquired myoelectric signals. These electrodes are preferred because
noninvasive techniques are much easier to access. In conventional myo-
prostheses (prostheses with electromyography techniques), two bipolar
EMG-signals are placed on antagonistic muscle groups, such as the wrist
extensors and flexors, and are used to control the velocity of one DOF
(Mazumdar, 2004). However, there exist techniques to collect intramuscu-
lar signals (Weir et al., 2009; Al-Ajam et al., 2013) with implantable myo-
electric sensors. More recently, a surgical procedure known as Targeted