IEEE Technology and Society Magazine - December 2015 - 25

movements. The motion capture
system must be accurate enough
to capture movements performed
by patients suffering from various
degrees of motion impairment, and
it should be suited for rehabilitation
training environments with no or
reduced medical supervision, such
as a patient's home or a community center. Indeed, several studies
have demonstrated that machine
learning systems are useful in biomedical signal processing as, for
example, for grouping among gait
variables or for multifunctional
prosthesis control [9].
A mong t he most com mon
approaches are different technologies that have been developed in
recent years to capture a person's
movement. To reveal and classify among the different kinds of
motions, different technologies
are commonly adopted to capture
upper extremity kinematics. Current technologies allow us to carry
out this task using several methodologies. It is possible to distinguish
between two broad categories of
motion capture methods: optical systems (equipped or not with
markers) and non-optical systems
(electromechanical, inertial, magnetic, and acoustic). A comparison
of technological and clinical pros
and cons regarding the most innovative systems adopted for TRH is
summarized in Table I:
■

least two cameras, requiring a
suitable calibration task. The
accuracy of the result depends
on the kind of markers used.
Non-optical systems use several
types of sensors for measuring
biometric and motion data.
■ Electromyography (EMG) systems are able to extract the
peripheral source that generates tremor, i.e., the muscles,
by recording the electrical
manifestation of muscular contraction, thanks to the use of
needles or surface electrodes.
■ Inertial measurement unit (IMU)
based motion tracking systems
are commonly considered as
most suitable for TRH applications, due to their advantages of
being compact, cost-effective, and
relatively easy to operate compared to their counterparts such
as visual-based tracking systems
[10]. They are able to capture the
movement of a person by placing
acceleration and angular rate sensors on the body surface.
■ RGB cameras, depth sensors,
and their combination called
"RGB-D sensors" present the
advantage of a higher spatial
resolution, although they can be
affected by the noise introduced
by obstruction and/or problems
relevant to the angle of view.
They have made possible the
development of an uncountable
number of applications in the
field of human-computer interaction (HCI). Such applications,
varying from gaming to medical,
are useful because of the capability of such sensors of elaborating depth maps of the placed
ambient. Many researchers have
focused on designing RGB-D cameras for biomedical applications.

Optical systems rely on infrared cameras that capture the
scene where the subject is moving. In marker-based optical systems, the subject is provided
with suitable markers capable
of reflecting infrared radiation
appearing as bright spots in the
scene. The resulting image is
a two-dimensional projection
of a three-dimensional scene,
which can be reconstructed by
means of triangulation from at
december 2015

In recent years, many devices
have been used as an alternative
tool to capture human movement.
∕

IEEE Technology and Society Magazine

In particular, there are some lowcost approaches to TRH by means
of the Microsoft Kinect sensor,
which not only has changed the
meaning of controller in computer
gaming, but it is also transforming the way we interact with every
technological device [11]. The possibility of rehabilitating two young
adults with motor impairments
using a Kinect-based system in
a public school setting has been
investigated in [12]. An interactive
game-based rehabilitation tool
for balance training of adults with
neurological injury has been developed and assessed in [13], while
the authors of [14] applied Kinect
to the problem of passive fall risk
assessment in home environments.
The use of Kinect and other natural gaming systems for developing
effective tools to aid rehabilitation
in clinical settings has been also
focused in [15]. The concurrent
validity of Kinect against a benchmark reference, a multiple-camera
3D motion analysis system in 20
healthy subjects during three postural control tests, has also been
assessed in [16].
A typical integrated system composed by innovative technologies
able to capture the movement in
a remote context is illustrated in
Figure 1. The patient needs to have
installed at home a monitoring and
detection system based on one or
more RGB-D devices and wearable
IMU sensors. Thanks to these devices, the data processing system is
able to track the patient's movements, their clinical status and the
degree of rehabilitation, without the
direct in situ supervision of a doctor or physiotherapist. A network
gateway is used to transmit the
sampled information and the doctor
will be able to see the data remotely, using a user-friendly interface
that eases reading and should allow
for a quick and easy interpretation

Table of Contents for the Digital Edition of IEEE Technology and Society Magazine - December 2015