Page 342 - Handbook of Biomechatronics

Upper Extremity Rehabilitation Robots: A Survey 335

The effectiveness of the FES therapy seems to be tied to the simultaneous
activation of sensory and motor pathways in the nervous system, which
coupled with the associated mental effort may increase the neuroplasticity
(Daly and Wolpaw, 2008). Therefore, the use of EEG in the detection of
motor imagery and proper timing of FES signals is proposed as a possible
solution to further improve the therapy outcome (Marquez-Chin
et al., 2016).

7.3 EMG-Based Strategies for Control and Rehabilitation
EMG signals are used to evaluate the amount of muscle activity during a spe-
cific task. If upper extremity rehabilitation robots target deficits in muscle
activations, their therapy will be more beneficial. The best way to capture
muscle activation patterns is to use bio-feedback (i.e., EMG) signals. In a
study by the Rehabilitation Institute of Chicago (RIC, United States), a spe-
cial voice and EMG-driven mobile exoskeleton (called “VAEDA glove”)
for hand rehabilitation has been developed (Thielbar et al., 2017). Com-
pared to other hand rehabilitation robots, the “VAEDA glove” is advanta-
geous since it allows for practice of functional task.
Poststroke patients were divided into two groups: (1) with rehabilitation
robot therapy (VAEDA) and (2) traditional fine-motor rehabilitation ther-
apy (No-VAEDA). The therapy was focused on grasp-and-release tasks. In
VAEDA therapy, the voice commands triggered the movement and the
EMG command drove the actuators. Results of this study showed that
the patients with VAEDA therapy could achieve better performances in
physiotherapy assessments.
Despite the satisfactory outcomes of EMG-based rehabilitation, it is not
suitable for performing complex movements. The success of EMG-based
methods highly depends on how well muscle synergies and activation pat-
terns are identified. The learning algorithm which is used to relate muscle
activations to physical activities plays an important role in the establishment
of better EMG-based rehabilitation. Advancements in deep learning will
provide a platform for EMG-based therapy in complex activities.

7.4 Model-Based Strategies for Control and Rehabilitation

Best design practices demand a proper understanding of the whole system,
which for this case consists of a human body interacting with a rehabilitation
robot. This interaction will affect rehabilitation procedures; however,
there is a lack of studies considering human body interaction with the