6B-5-4 
Axis Z 
: Ideal Initialization 
: Actual initialization 
Axis X 
Ü. 
Axis X' 
Figure 4-4 Difference of coordinate system 
-15 -10 -5 0 5 10 15 
E-W direction distance[m] 
■X—Traverse point —■—TEST1 
—TEST2 —♦— TEST3 
-•— TEST4 —©— TEST5 
Figure 4-5 Result of traverse survey 
4.2.2 Leveling(measurement of difference of elevation ) 
In this figure, the traverse point is a value measured with 
the automatic level. The gap (error) with the survey point is 
larger than that of the traverse survey. Here, the amount of 
the movement of the vertical direction to the movement 
time is smaller compared with the amount of the move 
ment of horizontal direction to the movement time. There 
fore, it is thought that the error of the leveling is larger than 
that of the traverse survey. Figure 4-6 shows the result of 
the movement tracks. Movement was in order as A—>B—> 
C-»D—>A. When moving from point A to B, actual move 
ment leaves at point A, then at once descends through the 
difference of elevation of about 30cm, and moves at a 
smooth place up to point B afterwards. According to tracks 
by the inertial survey of Figure 4-6, it is closer to actual 
movement in the place descending by about 30cm, but the 
part where it moves at a smooth place until reaching point 
B accumulates the error, and the tracks are quite different 
from actual movement. 
-10 0 10 20 30 40 
N-S direction distancefm] 
X Traverse point 
■ 
-TESTI 
-A-TEST2 
-TEST3 
—TEST4 
—e- 
-TEST5 
Figure 4-6 Leveling (movement tracks) 
The accumulation of the error is a little in TEST4 and 
TEST5, because a straight line distance is short, even 
though the movement from point B to C is of an almost 
smooth place. 
In the movement from point C to D, it actually leaves at 
point C and moves at a smooth place, then goes up by 
about 30cm, and moves some meters to reach point D. 
According to tracks of Figure 8 we are supposed to move 
at a smooth place but as the error is accumulated, the 
tracks are descended. However, the tracks are showing 
almost an actual motion only for the place going up. 
The movement from point D to A' (point which returned to 
point A) is the same as the one from point B to C, and 
moved distance is of the straight line is short and the ac 
cumulation of the error is small. 
From the above-mentioned point of view, it is thought that 
if the place with some difference of elevation is moved 
without taking much time in the leveling by the inertial 
survey, a value close to an actual one can be calculated. 
However, it is understood that the error accumulates when 
time is taken for measurement at the place where there 
are not so many differences of elevation and a preferable 
result is not obtained.