Biomedical Engineering Reference
In-Depth Information
of a specially developed calibration device, called “A-palp”, allowing the observer
to use the finger tip directly as digitizing tool without requiring the interposition of
cumbersome locating device [
50
] (Fig.
9.3
e-g).
Once collected, motion data must be represented using some conventions. The
ShRm is frequently reported following the Euler-Cardan convention [
10
,
12
,
18
,
20
,
35
,
45
,
51
-
59
]. Results in the present chapter will also be represented using the Car-
dan ZXY convention, next to attitude vector (orientation vector, following Cappozzo
[
60
]) projections. The use of attitude vectors and related helical axis representation
for rotational degrees of freedom (DoFs) is justified because such procedures have
been intensively reported as the most robust methods for general joint motion repre-
sentation [
13
,
40
,
47
,
48
,
60
-
71
]. FollowingWoltring [
71
], such approach of motion
data storing seems robust for segment motion evaluation in global coordinate sys-
tem: “
The attitude vector dispenses with the 'gimbal-lock' and non-orthogonality
disadvantages of Cardanic/Eulerian conventions; therefore, its components have
better metrical properties, and they are less sensitive to measurement errors and to
coordinate system uncertainties than Cardanic/Eulerian angles.
”
9.2 Data Collection for Scaling Methods
This section describes the specificities of the data collection performed to determine
the ShRm, keeping in mind the modeling context of the model-based approach to be
developed. The main aim is to find a good compromise between anatomical reality
and the constraints of in-vivo modeling (e.g., within an ergonomic context or for
musculoskeletal analysis).
The main challenge of the underlying research is to find on the one hand the
relationships between the humerus position and orientation and on the other hand
the related attitude of the scapula and clavicle at the same moment of time. This
answers a practical problem in in-vivo motion analysis. Indeed, although data related
to the humerus instantaneous spatial position is relatively easy to obtain (for example,
using stereophotogrammetry based on reflective markers), the same information is
more difficult to collect for the clavicle (which is a small elongated bone offering
limited space to attach markers) and the scapula (of which the real motion is largely
hidden beneath the shoulder soft tissue). Furthermore, a unique and straightforward
one degree-of-freedom mechanism like in the knee joint or the ankle joint [
72
]is
physiologically not observable (this is depicted in Fig.
9.3
a-d).
This is due to the fact that motions within the sternoclavicular joint will be similar
during shoulder translation (that does not require displacement within the scapulo-
humeral joint) and scapulohumeral abduction. Consequently, it is not directly pos-
sible to estimate clavicle instantaneous orientation from one unique humerus pose.
This is clearly visible in Fig.
9.3
b: both clavicles show similar angles with the thorax
line, while the humerus orientation is very different. The scapula shows similar rela-
tionships with the clavicle through the acromioclavicular joint and with the thorax
through the scapulothoracic joint. The above example demonstrates that an algorithm
