In cace of normal walking, case A, both knee of 
most people had approximately smooth sine 
curves, different in their slopes in spatial 
displacement. 
In case of experimentally stressed walking, case 
B, ie. a right knee and ankle fixed by a knee- 
ankle-foot plastic medical orthosis, the knees 
showed a sine-curve with small slope and 
amplitude and with short swinging phases of the 
stressed leg. 
But a damaged leg of some persons (case C), who 
had experienced a broken leg or knee ligament had 
a sine curves with a small and amplitude and with 
short supporting phases. The normal leg, which is 
stressed by problem to the other leg, had a similar 
curve and phase as that of the problem leg. 
The result of sum of metabolic energy is as 
follows: 
Case A: normal rt leg = normal It leg 
Case B: normal It leg > stressed rt leg 
Case C: normal rt leg> damaged It leg 
It is said that normal leg has burdens by 
influences of leg problems. 
log PS 
Case A 
  
log PS 
Case B 
stressed rt knee 
  
  
  
  
" 
Fig. 6 Power Spectrum of Knee (X-axis) He 
C. Spectral Analysis 
Spectral analysis suggested two features. Fig. 6 
shows power spectrum of knee (X-axis), curves of log 
PS vs Hz indicate new information for features of 
walking. 
In case of natural normal (Case A) and stressed 
walking (Case B), most people had a similar power 
spectrum curve for the right and left knee. 
The slope of that power spectrum curve was 
proportioned to approximately, i.e.. power spectrum 
was inversely proportioned to frequency. But those 
(Case C), who had experienced a broken leg or knee 
ligament had quite different power spectrum curves 
for the right and left leg. 
The normal leg had 
1/ f* proportion, 
the damaged leg had frequencies higher than the 
those of normal leg. 
The 1/f* proportion 
is depended on person's movement. 
CONCLUSION 
Mechanisms of movement have many factors, it is 
important to approach to analyze from the view of 
Biomedical Engineering using 3-D videogrametry. In 
the further study, it will be discussed about the 
feedback control relationship between right and 
left knees in walking. Analysis of Impulse Response 
utilizing AR modeling will indicate mechanisms of 
movement. 
ACKNOWLEDGMENT 
The authors would like to thank for experimental 
supporting by many students in Institute of Industrial 
Science, Univ. of Tokyo. 
REFERENCES 
[1] H. Akaike (1974), "A new look of the statistical 
model identification" , IEEE Trans. Automat. Com. 
vol. 19 no.6, pp. 716-722 
(2] M. B. Priestley (1984),"Spectral Analysis and 
Time Series", Academic Press. 
[3] F. C. T. van der Helm (1995)," Three-Dimensional 
Recording and Description of Motions the Shoulder 
570 
International Archives of Photogrammetry and Remote Sensing. Vol. XXXI, Part B5. Vienna 1996 
Mecha 
Engine 
(4] R 
"Synth 
Functi 
Trans. 
pp. 67- 
[5] G 
Digit: 
Mover 
Proce: 
Engine 
[6] M. 
and 
Mathe 
Medic 
data f 
Senso! 
Symp 
Photo; 
pp.72 
[7] M. 
and T 
of al 
CCD 
Interc 
Seque