Biomedical Engineering Reference
In-Depth Information
100
90
80
70
60
50
Experiment
Prediction
40
30
20
10
0
1
2
3
4
5
6
7
8
9
10
11
12
Participants
FIGURE 9.15
Normalized rating scores of 12 participants evaluating the predicted and experimental
box-lifting task.
acceleration (
Figure 9.17B
) shows behavior similar to that of the velocity; how-
ever, more local fluctuations appeared in the experimental curves.
For the two important joints for this task (hip and torso), the calculated veloc-
ity and acceleration profiles are shown as compared with values obtained from
the experiments (
Figure 9.17
). It is shown that the PD method yields adequate
results and is indeed a rigorous approach to predicting the motion.
For kinetics,
Figure 9.18
shows the experimental and predicted vertical and
forward GRF during lifting. As indicated, the predicted model showed behaviors
similar to those of the experimental data but was not able to capture the initial
characteristics of the lifting cycle.
9.6.4.2
Quantitative comparison
The comparison between the experimental and the predicted lifting determinants
in terms of the interval of confidence is presented in
Figure 9.19
. The experimen-
tal data for each subject represents the average of two lifting cycles. The model
determinants show weak correlation for the ankle and the shoulder flexion; how-
ever, they show reasonable correlation and agreement with the human subjects for
the rest of the determinants, by being inside the interval of confidence and follow-
ing the trajectories of the subjects' determinants.
Figure 9.20
depicts the R
2
plot
for the six lifting determinants for the distinctive key frames (0% height, 20%
height, and 40% height). The solid black circle in
Figure 9.20
represents the rela-
tionship between the simulation data and the average experimental data of the hip
Search WWH ::
Custom Search