Gyros
Gyros
Gyros
Detection range 90 s
Responsibility DC ~ 7Hz
Voltage 8 ~ 13.5V DC + +
Power sup.
Current 15mA <
Weight 45gf <
Item Specification
Resolution 0.06 s
Detection range 90 s
Responsibility DC ~ 20Hz
Voltage 12V DC
Power sup.
Current 1A <
Table 1. Specification of accelerometer JA-5V
Figure 4. Running through chart of the experiment place
Figure 1. Coordinate system and sensors
Axis-Z
Axis-Y
Axis-X
Accelerometer
Gyro
Axis-Z
Axis-Y
Axis-X
Accelerometer
Gyro
Figure 2. Experimental equipment
Table 2. Specification of Vibration gyro ENV-05A
Table 3. Specification of fiber optical gyro JG-35FD
Figure 3. Plan of experiment place
Figure 4. Running through chart
3
3.2 Fixed and moving coordinate systems
Two kinds of coordinate system are used in the
experiments: the one denoted by ( ) XYZ is fixed to the
ground such that X-Y plane is horizontal and X-axis
directs toward magnetic north, and the other one denoted
by ( ) xyz moves with the equipment. They shall be
called the fixed coordinate system and the moving
coordinate system, respectively. Both systems coincide
with each other at the beginning of the experiment, but
the system ( ) xyz translates and rotates as the inertial
navigation system goes along the route. Accelerations
and angular velocities obtained from the sensors are the
vector components in the moving coordinate system
( ) xyz . Therefore, they are transformed into those in the
fixed coordinate system ( ) XYZ before integration, and
the rest of the computations (e.g. positioning and error
evaluation) are performed in ( ) XYZ .
3.3 Effects of the moving speed
In inertial survey, the distance is obtained by integrating
the acceleration twice, and the angular change is
obtained by integrating the angular velocity once. The
error accumulates and increases with the lapse of time,
and hence it seems desirable to finish the measurement in
shorter time to make the positioning more accurate. In
order to ascertain whether this inference is correct or not,
experiments were repeated changing the moving speed of
the equipment. The nonstop strategy is adopted in this
experiment with three kinds of moving speed: slow
(approximately 3 min. to go around whole the route,
approximately 0.9 m/s), medium (2 min., 1.4 m/s) and
fast (1 min., 2.7 m/s).
Figure 5(a) shows the estimated trajectories by the
inertial navigation system equipped with fiber optical
gyros. They are projected onto a horizontal plane. Thick
solid curve represents the true trajectory mentioned in
section 3.1, and other three curves stand for the estimated
trajectories in cases that the moving speed is slow,
medium and fast, respectively. As has been expected, the
discrepancy between the true and the estimated
trajectories decreases as the moving speed increases. If
Figure 5. Estimated trajectories projected onto
a horizontal plane (nonstop strategy)
Fiber optical gyro
-40
-20
0
20
40
60
80
100
120
140
-250 -200 -150 -100 -50 0 50
Y[East-West][m]
X
[
N
o
r
t
h
-
S
o
u
t
h
]
[
m
]
True value 0.9[m/s] 1.4[m/s] 2.7[m/s]
Vibration gyro
-100
-50
0
50
100
150
200
250
300
350
-600 -400 -200 0 200
Y[East-West][m]
X
[
N
o
r
t
h
-
S
o
u
t
h
]
[
m
]
True value 0.9[m/s] 1.4[m/s] 2.7[m/s]
( ) a
( ) b
Fiberoptical gyro
0
50
100
150
200
0 1 2 3
Moving speed[m/s]
H
o
r
i
z
o
n
t
a
l
D
i
s
c
r
e
p
a
n
c
y
[
m
]
0.9[m/s] 1.4[m/s] 2.7[m/s] Approximate curve
Vibration gyro
0
200
400
600
800
1000
0 1 2 3
Moving speed[m/s]
H
o
r
i
z
o
n
t
a
l
D
i
s
c
r
e
p
a
n
c
y
[
m
]
0.9[m/s] 1.4[m/s] 2.7[m/s] Approximate curve
( ) a
( ) b
Figure 6. Dependence of the horizontal discrepancy
on the moving speed (nonstop strategy)
4
the fiber optical gyros are replaced with the vibration
gyros, we get the results shown in figure 5(b). Since the
resolution of the vibration gyro is lower than that of the
fiber optical gyro, the discrepancy is larger in figure 5(b)
than that in 5(a). In order to evaluate the discrepancy
quantitatively, the horizontal discrepancy
h
is
defined as the horizontal distance between the true
position and the estimated one at the end point of the
route, namely
2 2 1 2
(( ) ( ) )
h e t e t
X X Y Y = + where the
subscripts e and t refer to the estimated and the true
coordinates of the end point, respectively. The
dependence of the horizontal discrepancy on the moving
speed is plotted in figure 6. The experiments were
repeated three times for each moving speed, and hence
three points are plotted for each moving speed in both
panels of figure 6.
3.4 Effects of the stop-and-go strategy
The stop-and-go strategy was proposed in order to
suppress accumulation of integration errors due to
random drift of the accelerometers and gyros. In this
strategy, the equipment moves from the starting point
0
R and halts at a point
1
R on the route, then it moves
from
1
R and halts at the next point
2
R on the route, and
so on. These procedures are repeated until the equipment
reaches the end point
N
R . Suppose that the equipment
leaves from
1 n
R at time point
n
t and gets to
n
R at
n
T ,
which means that the equipment moves during the time
interval [ , ]
n n
t T for 1 ~ n N = and stays still during
1
[ , ]
n n
T t
+
for 1 ~ 1 n N = . Suppose more that the
integration of the acceleration ( ) t measured by the
accelerometers over the interval [ , ]
n n
t T is
n
v , which
would be 0 if the measured acceleration were free from
noise. Then, in order to make the velocity at
n
T zero,
( ) t is replaced with ( ) ( )
n n n
t T t v in the interval
[ , ]
n n
t T . This method of compensation is known as
ZUPT (Zero-velocity Up Dates). Anyway, in order to
make this method effective, it is crucial to record the
timings ,
n n
t T accurately.
In the actual experiment performed this time,
the stopping points ( 0 ~ )
n
n N = R are arranged
uniformly over the route with the distance between the
adjacent points, or the stoppage interval, about 3, 5, 10,
20, 30, and 50m. This means that the experiment was
repeated 6 times with these 6 choices of stoppage
interval. The velocity of the equipment was kept as
constant as possible to about 1.4 m/s that is referred as
medium speed in the previous section, and the length
of the time interval
1
[ , ]
n n
T t
+
or the stay time is chosen to
be 5s. In each panel of figure 7, the true trajectory is
plotted with a thick solid line as the same in figure 5,
while a thin solid curve represents the estimated one by
means of the stop-and-go strategy in the case that the
stoppage interval is 10m. Moreover, the estimated
trajectory by means of the nonstop strategy with medium
moving speed is also plotted with a thin dotted line as a
comparison. It is clear that the estimation error is
reduced drastically by means of the stop-and-go strategy.
4 BIAS CORRECTION
When the accelerometers and gyros stay still, their
outputs should be zero. However, in reality, some
nonzero output called the bias error is observed even
form a sensor at a standstill as described in figure 8. The
bias error
B
in the acceleration yields the positioning
error
2
2
B
t where t is the elapsed time, while the bias
error
B
in the angular velocity causes the angle error
B
t . The magnitudes of bias errors observed from the
sensors used in the present experiments are listed in table
4. The positioning and angle errors due to these bias
errors are plotted in figure 9 as functions of the elapsed
time. As can be seen in this figure, the positioning errors
become 16, 70 and 6m along X, Y and Z axes,
respectively, when the elapsed time is 60s. In the case of
fiber optical gyros, the angle errors 60s after the onset
are 3.2, 1.1 and 0.3 degrees around X, Y and Z axes,
respectively, while they are as large as 47, 38 and 5
degrees in the case of vibration gyros.
Figure 7. Trajectories estimated by means of
stop-and-go and nonstop strategies
Fiberoptical gyro
-20
-10
0
10
20
30
40
50
60
70
-250 -200 -150 -100 -50 0 50
Y[East-West][m]
X
[
N
o
r
t
h
-
S
o
u
t
h
]
[
m
]
True value Stop&Go Non-Stop
Vibration gyro
-50
0
50
100
150
200
-400 -300 -200 -100 0 100
Y[East-West][m]
X
[
N
o
r
t
h
-
S
o
u
t
h
]
[
m
]
True value Stop&Go Non-Stop
( ) a
( ) b
5
The bias errors are estimated at each stopping point
1
( 1 ~ )
n
n N
16m
Y-axis
70m
Z-axis
-6m
Fiber optical gyro
0
1
2
3
4
0 20 40 60
Time[s]
E
r
r
o
r
i
n
a
n
g
l
e
[
d
e
g
]
X-axis
3.2deg
Y-axis
1.1deg
Z-axis
0.3deg
Vibration gyro
0
10
20
30
40
50
0 20 40 60
Time[s]
E
r
r
o
r
i
n
a
n
g
l
e
[
d
e
g
]
Y-axis
38deg
X-axis
47deg
Z-axis
5deg
( ) a
( ) b
( ) c
Figure 10. Effects of the bias correction applied to
the inertial survey together with the stop-
and-go strategy.
Vertical alignment
-6.0
-5.0
-4.0
-3.0
-2.0
-1.0
0.0
1.0
0 50 100 150 200
Horizontal distance[m]
E
l
e
v
a
t
i
o
n
D
i
s
c
r
e
p
a
n
c
y
[
m
]
True value Before correction After correction
Plane alignment
-20
-10
0
10
20
30
40
-140 -120 -100 -80 -60 -40 -20 0 20
Y(East-West)[m]
X
(
N
o
r
t
h
-
S
o
u
t
h
)
[
m
]
True value Before correction After correction
( ) a
( ) b
6
5 CONCLUSIONS
Conclusions are as follows:
(1) The stop-and-go strategy proposed by the authors
makes the inertial survey more accurate than the
conventional nonstop strategy.
(2) The accuracy increases as the speed of the equipment
increases.
(3) The accuracy is strongly influenced by the perform-
ance of gyros.
(4) The accuracy is improved by the stop-and-go strategy
even with low-performance gyros.
(5) The bias correction is quite effective, especially when
it is used together with the stop-and-go strategy, to
improve the accuracy.
REFERENCES
1) Innes J.: Inertial platform techniques for rapid surveys
of hilliness and bendiness, Sensors Highway and Civil
Engineering; London, p.143-154, 1981
2) Chapman W. H.: Proposed specifications for inertial
surveying, Tech. Paper Annu. Meeting American
Congr. Survey Mapping, No.43, p.287-293, 1983
3) Roof E. F.: Inertial survey systems, J. Surv. Eng
Vol.109, No.2, p.116-135, 1983
4) Cross P. A.: Inertial surveying, Int. J. Remote Sensing,
Vol.6, No.10, p.1585-1589, 1985
5) Suzuki H. et al.: Real-time measurement of altitude
data using the accelerometer, International Archives
of Photogrammetry and Remote Sensing, Hakodate,
Vol. XXXII, Part 5, p.278-283, 1998
6) Koizumi T. et al.: Fundamental study on development
and application of the local positioning system using
accelerometer and gyroscope, Proc. Int. Workshop on
Mobile Mapping Technology, Bangkok 6B-5-1 - 6B-
5-6, 1999
7) Shirai Y. et al.: Fundamental study on real time
measurement of altitude data with accelerometer and
vehicle speed sensor, International Archives of
Photogrammetry and Remote Sensing, Amsterdam,
Vol. XXXIII, Part B1, p.301-306, 2000