Xu Bo Sun Feng
College of Automation College of Automation
Harbin Engineering University Harbin Engineering University
Harbin,China Harbin,China
xubocarter@hotmail.com
tlj119911@yahoo.com
Abstract—With the development of fiber-optic gyroscopes,
strapdown inertial navigation system has become a trend in the
development of inertial navigation system (INS), and
strapdown inertial system error calibration technique has
become one of the key technologies. This paper combined
engineering practice with study an on-line calibration method
of INS, which based on the high-precision multi-functional
three-axis turntable. First of all, aim at the actual system used
in fiber-optic gyroscopes, accelerometers quartz flexible
components, set up corresponding system error model
respectively. Then, take the high-precision three-axis
turntable as platform to design the calibrate path, through the
on-line combination calibration .Use Kalman filter algorithm
to process data on-line, completed the corresponding
experiments, and verified by tests. Experimental results show
that the calibration method is feasible in project and effective,
and increased level of automation.
Keywors :Online Calibration; Three-axis Turntable; INS
I. INTRODUCTION
For INS, error compensation is an effective way to
improve the INS navigation accuracy, and the error
calibration is a prerequisite to error compensation. INS
calibration can be divided into system-level calibration and
separation calibration according to the difference of
measurement. Separation calibration is the classic method,
which is also abroad used at present. Conventional
calibration methods have some shortcomings as follows: (l)
Processing afterwards, real-time performance is not fine. (2)
Large amount of data, which required recording more data,
and along with the calibration parameters increased the
amount of data leap and time-consuming .Calibration
accuracy depends directly on the accuracy of the turntable
accuracy.
System-level calibration method is developed since the
beginning of the 1980s, which represents an important
developed direction of calibration technology. This method
is used when the navigation system come into the state of
navigation. It takes navigation error (including the error of
location, the velocity error and attitude error) as
measurement to calibrate the system error parameters. On
static base conditions, the calculated value of location and
velocity from navigation is the location, velocity error. In the
navigation state, the error of the gyroscope through the error
transfer affected the output of the system, that is to say, the
system output includes information of all kinds of error
source. So take the system output as measurement can be
used to estimate the error parameters of the system.
System-level calibration method points out to use
navigation error to identify IMU error, there is literature says
that use velocity error as measurement, which use dual-axis
position turntable to calibrate inertial system error
parameters. The advantage of the method is not need
high-precision turntable, but only use velocity error as
measurement, but some parameters observability is
poor .Results of the calibration are also poor, and rotation of
multi-location program operation is complex, calibration is
time-consuming. As a result, new calibration method which
research based on high-precision three-position rate
turntable is very necessary. This paper bases on the
three-axis turntable in our laboratory, study the on-line
calibration method which adapt attitude error and angle error
as measurement, respectively do the estimation and analysis
on gyroscope model error.
II. G
YRO ERROR MATHEMATICAL MODEL [2] [3] [11]
On the assumption that gyroscope output is
g
N .If there
is no scale factor error, bias and installation error, and other
random drift error, we can get the angular velocity input
from the gyroscope output:
b
ib g
K N
ω
=⋅ (1)
Due to the existence of error, the actual angular velocity
input is
()()
bb b
ib g g g g
bb
ggggg
bCKN
bICKISN
ωδε
δε
=+ +
=++Δ + +
(2)
After ignored small amount of second-order
()
()
bb b
ib g g g g
bb
ggggg
bb b
gggibg
bCKN
bIS CKN
bIS C
ωδε
δε
ωδε
=+ +
=++ +Δ +
=++ +Δ +
(3)
Gyro output error is
()
bbbb
gg gib
bS C
εωδε
=+ +Δ + (4)
In which:
[]
bbbbT
xyz
εεεε
= is the gyro output error;
A FOG Online Calibration Research Based on High-precision
thr
ee-axis Turntable
2009 International Asia Conference on Informatics in Control, Automation and Robotics
978-0-7695-3519-7/09 $25.00 © 2009 IEEE
DOI 10.1109/CAR.2009.48
454