Digital Signal Processing Reference
In-Depth Information
Tabl e 2
Effect of 1% scale factor error in accelerometer to functions of signal
patterns for step length estimation
Raw
Square root
Cube root
Fourth root
1
/
2
1
/
3
1
/
4
Function
p
j
(
k
)
p
j
(
k
)
p
j
(
k
)
p
j
(
k
)
Step length error (%)
1.00
0.50
0.33
0.25
a
b
c
Heading error (degrees)
y−c
oordina
te (m)
True
Est
R
e
lative position error (% of distance traveled)
4
4
30
3
3
20
2
2
10
1
1
0
0
0
0
2
4
6
8
10
0
200
400
600
800
0
2
4
6
8
10
Time (min)
x−coordinate (m)
Time (min)
Fig. 8
Effect of 25
◦
/
h gyro bias when the pedestrian is walking with constant speed along the
positive
x
-axis: (
a
) heading error, (
b
) true and estimated coordinates, and (
c
) relative position error
Tabl e 3
Comparison of gyro grade with respect to the effect of uncompensated
bias to the PDR error build-up
Navigation
Tactical
Consumer
Bias instability (
◦
/
h)
0.0035
1
25
Time to 2% relative position error
27 days
2.3 h
5.5 min
Time to 3% relative position error
41 days
3.4 h
8.3 min
Time to 3
◦
heading error
35 days
3.0 h
7.2 min
the long term average in the relative positioning error is about 2-3%, as can be
introduced by the gyro bias is smaller in the beginning, but exceeds 2% in less than
6 min and 3% in less than 9 min.
with gyro instabilities typical to navigation and tactical grade gyros. The results are
The effect of the gyro scale factor error is simulated by using a scenario where
the pedestrian first makes a 180
◦
turn and then walks with a constant step length of
0
.
75 m and a constant frequency of 2 steps/s along the positive
x
-axis. The gyro scale
4
The growth is almost quadratic with small heading errors; however, with larger heading errors,