Page 201 - Handbook of Biomechatronics

198 Georgios A. Bertos and Evangelos G. Papadopoulos

actuator dynamics acting through some physical mechanical compliance,
aiming at implementing variable impedance according to the task to be per-
formed, (Vanderborght et al., 2013).
An important issue to all control types is the type of command they
accept and its interpretation. In the dominant in the early- to mid-1900s
classic extended physiological proprioception (Classic EPP), no controller
was needed and the connection between the end effector (implement)
and the remaining limb was purely mechanical: the prosthetic limb was con-
nected directly to cineplasty sites of residual arm with Bowden Cables
(Tavakoli et al., 2017; Klopsteg and Wilson, 1954). As EPP was abandoned
in favor of electromechanical prostheses, EMG was used as a high-level
command to the active limb.
An EMG-based control system or myoelectric control system, controls
the limbs by converting muscle movements to electrical signals allowing the
amputees to control the prosthesis more directly (Harvey and Masland,
1941; Jawhar et al., 2011). The EMG signals must be amplified, filtered,
and processed to yield root-mean-square (RMS) signals appropriate as con-
trol references. However, this introduces undesirable delays in the motion of
the prosthetic limb. Unlike to EPP, this type of control does not provide
feedback to the patient (proprioception) even if internal feedback is pro-
vided to the actuators for low-level control; its implementation requires
visual feedback (Cloutier and Yang, 2013). An overview of myoelectric
control, and its performance with respect to the characteristics of the ideal
myocontroller is presented (Farina et al., 2014). Classic and relatively novel
academic methods are described, including techniques for simultaneous and
proportional control of multiple-DoF, and the use of individual motor neu-
ron spike trains for direct control. Although myoelectric signals are widely
considered as the best available control interface for powered prostheses,
many amputees abandon their devices out of frustration due to the lack
of precision of the prosthesis’ movements (Shehata et al., 2017).
Ideally, a prosthetic device should establish a bidirectional communication
between the patient and the prostheses. Currently and in principle, two
methods can provide bidirectionality, the Biomechatronic EPP (Mablekos-
Alexiou et al., 2015; Moutopoulou et al., 2015) and the direct neural inter-
faces (invasive or not) (Di Pino et al., 2009; Jerbi et al., 2011). The former
represents new topology (Fig. 13) of EPP and aims at elimination of the
drawback of cineplasty and Bowden cables, which render the EPP unaesthetic
for the user. The core of this concept is based on principles of the field of
telerobotics and teleoperation (Yokokohji and Yoshikawa, 1994). In this