Biomedical Engineering Reference
In-Depth Information
9.3.2 Motion Representations: Comparison of Basic Motions
In this section we present the comparison of the ShRm evaluation using Cardan
(ZXY) and OVP (Orientation Vector Projected on proximal LCS as explained above)
conventions. Then, the evaluated ShRm will be illustrated by the model-based
shoulder motion simulation in near scapula and frontal plane humerus elevations.
9.4 ShRm Regression Equations
This section contains results related to the estimation of the position and orientation
of the clavicle and scapula from the humerus attitude.
Figure
9.8
shows typical raw source data and results from regression as obtained
for all DoFs processed in this work. Raw source data (plotted by colored dots)
together with results of fitting (full curves, equations are given in the legend box)
are given for one DoF (flexion-extension). This figure shows displacements for
the right clavicle (longitudinal rotation) and right scapula (tilt motion). The defi-
nitions of the DoFs are as explained in the first section of this chapter. This figure
shows linear and parabolic polynomial fitting of the right sternoclavicular (Clav,
[
00052921x
2
0
.
+
0
.
20886x
+
13
.
7866; mean error
=
1
.
5
(
SD
=
1
.
1
)
]) and acromio-
00043213x
2
clavicular (Scap,
[
0
.
+
0
.
10288x
−
15
.
5552; mean error
=
1
.
1
(
SD
=
0
]) motions compared to the humerus sagittal displacement in the thoracohumeral
joint (flexion/extension) using the ZXY conventions. Precision of fitting is also indi-
cated in the figure caption behind each fitting equation (with mean and standard
deviation).
Results for the other DoFs are given in Tables
9.5
and
9.6
for the clavicle and
scapula, respectively.
Figure
9.9
shows the results of the comparison between Cardan (ZXY, on the left)
and OVP (on the right) conventions for data representation. It is illustrated by the
fitting of clavicle (elevation) and scapula (upward rotation) motion around the X-axis
of the proximal body versus the humerus thoracohumeral abduction/adduction.
Comparison of the fitting residual (e.g., for the clavicle data: Cardan representa-
tion
.
9
)
) shows that OVP conventions seem more
appropriate to approximate the scapula and clavicle behavior.
Tables
9.5
and
9.6
contain values of linear and parabolic fitting coefficients of
three rotational DoFs for clavicle and scapula respectively. Residual of fitting analy-
sis shows that both ZXYcardan andOVP conventions show similar mean(SD) values,
but comparison of clavicle motion for humeral abduction/adduction motion shows
lower residuals for the OVP approach (i.e., compare fitting residual in the mean(SD)
columns of Table
9.5
for XYZ lines from 7 to 12). Therefore, OVP convention was
used in this work for shoulder rhythm prediction and model-based simulation. Trans-
formation to ZXY cardan convention was then performed using a rotation matrix.
The analysis of the results presented above suggests that linear multiple regres-
sions must be applied for final clavicle and scapular orientation prediction (e.g.
=
2
.
1
(
2
.
1
)
and OVP
=
1
.
4
(
1
.
1
)
