Biomedical Engineering Reference
In-Depth Information
modalities [
78
]. All of these approaches tend to deliver various types of feedback
that are more or less linked to dynamic properties.
As no device exists to directly measure joint torques, this type of information
seems difficult to apply for navigation in virtual environments, except maybe for
evaluation purposes. However, muscle activation can be monitored thanks to elec-
tromyographic sensors and could thus be used either as an input signal or an output
for evaluation purposes. Indeed, electromyograms (EMG) (see Part 2, Chap.
3
“sens-
ing human walking” of this topic for an introduction of this measure) provide us with
muscle coordination or muscle activation patterns and can give interesting informa-
tion about walking performance. Figure
3.5
shows the muscle activation pattern of
the most relevant muscles when walking [
59
]. It clearly shows that despite large
inter-individual variations, a stereotyped shape appears for each muscle. It could
thus be possible to apply signal processing and pattern recognition to determine in
which phase of the walking cycle the subject is when activating the muscles. Some
authors in biomechanics have used such type of information to simulate phases of
a gait in musculoskeletal models [
33
]. Other authors were able to estimate the joint
angle velocities depending on EMG signals [
19
].
EMG is very difficult to measure with good accuracy and it is sometimes difficult
to correctly deduce the muscle activation. However, the main idea here could be
to analyze EMG as people do with Electroencephalography (EEG) signals used in
Brain Computer Interfaces [
40
]. We could imagine that EMG electrodes placed over
relevant muscles of the lower-limb may detect activation patterns that could be an
input for the navigation system. They can also be used for control if the proposed
walking interface generates natural muscle activation patterns when navigating in
virtual environments. However, it seems that this modality has not been used yet to
control navigation in immersive environments.
3.3.2 Energetics of Human Walking
Many researchers have shown that locomotion is a very economic motion from
the energetic point of view, especially for naturally selected speeds [
17
,
18
]. People
are able to walk for a long time and along very long distances. However, people
have to support and propel their body mass, which seems to be a costly task from
the energetic point of view. The reported minimal energy expenditure is thus the
result of a positive coordination pattern.
Firstly, there exists an anti-phase coordination between the global potential and
kinetic energies [
13
,
75
], which enables us to maintain an almost constant total
amount of energy without introducing huge mechanical work (see Fig.
3.6
).
The anti-phase coordination pattern between kinetic and potential energies favors
transfers between these two energies. This phenomenon can be observed in an in-
verted pendulum, which is a very good approximation of the human body during a
step, especially during casual walking [
13
]. This energy transfer is less efficient when
the gait parameters diverge from those observed during casual walking [
17
,
18
].
