meme
21017205)
Jones
eI
0017203)
mnm
06172058.)
listance
lane
istance in
vided at
hron tape
of large
y,and the
detection
) of the
er in the
ed to be
K ) in the
using a
pordinate
transformation was performed by using an expression
of X=(acceleration value of X-axis) X cos « to calculate
the plane coodinates.The accelerometer and vibration
gyro ware provided on the carriage,a plumb bob was
attached at the rear of the carriage for centering.
The measuring time was about 30 seconds for
10m,about 50seconds for 20m,and about 1 minute for
30m. The error for the moving distance of 10m was
1.32%(precision1/77),the error for the moving distance
of 20m was 1.75%(precision1/57),the error for the
moving distance of 30m was 2.56% (precision1/40).
3.3Distance measuring experiment in the height
direction
Next,in order to calculate a height.a leveling rod was
vertically installed,and an accelerometer was moved
along the leveling rod in the vertical direction.An
accelerometer installed on a block was mounted on a
pedestal which had been placed to be horizontal along
the leveling rod with a level, and was manually moved
up to a predetermined height along the graduation on
the leveling rod. The measuring time was about 10
seconds for 50cm,about 10 seconds for 100cm.and
about 15 seconds for 150cm.The error for 50cm was
0.60%(1/170),that for 100em was 1.28%(1/80),and that
for 150cm was 1.5396(1/70).
The accelerometer for use with this research always
outputs the gravitational acceleration,and when the
detection axis direction of the accerelometer is in the
vertical direction,a voltage of about 5V is outputted.
According to the experiment, it could be confirmed that
when an accelerometer,which always detects
gravitational acceleration,is used,and has been moved
in the vertical direction,the height along which it has
been moved can be calculated.It is smaller than the
above-described error when it has been moved in the
horizontal direction.As the causes therefor the
following may be considerd:(D there were few data
owing to the short measuring time,and the
accumulated error for the integrated values was small;
© there was less vibration(noise),etc.in the vertical
direction during the movement; and (9 no coodinate
transformation was performed because it was assumed
that it was moved in the gravitational direction.
4.Application to Survey
4.1Leveling experiment and consideration
In order to perform an experiment for calculating a
height and a moving distance by using an inertial
device,a wheel-barrow made of aluminium mounted
with an inertial device was moved.
For the experimental method,a difference in level of
0.469m in a longitudional level difference was provided
9
on a horizontal distance of 4.765m shown in Fig.6,and
the wheel-barrow mounted with an inertial device was
moved to compare with the leveling using a level.The
accelerometer in the X-axis direction,which is the
traveling direction, was turned toward the next survey
point,and was moved linearly between the survey
points maintaining it horizontal with a level mounted
on the wheel-barrow.As regards the survey points,a
plump bob was attached on one side of the tire of the
wheel-barrow for centering.
The accelerometers and the vibration gyros installed to
the three axes(See Fig.7).
The data used for the analysis are X-axis
accelerometer(Vx), X-axis vibration gyro( $ ),Y-axis
accelerometer(Vy),Y-axis vibration gyro( o ),Z-axis
accelerometer(Vz), Z-axis vibration gyro( « ).
Fig.D.Experinment place
Zaxis
Xaxis Xa Ya Za: ACCELEROMETER Y axis
Xa Ya Za: VIBRATION GYRO
Fig.7 Three axes and sensors
When rotary movement was performed by kx
clockwise with respect to the posotive direction of Z-
axis in an absolute coordinate system XYZ before
movement,next similarly by $ around the Y-axis,and
further by « around the X-axis in order,it is assumed
that it can be converted to an inclined device
coordinate system after the movement.
For simplification,as regards X-coodinate value,Y-
coodinate value,and Z-coodinate value in a device
coodinate system which has inclined when the
rotations by «. ¢ .and o were rotated independently
International Archives of Photogrammetry and Remote Sensing. Vol. XXXI, Part B1. Vienna 1996