GPS-Aided Calibration Method of Gyro Parameters
Date Issued
2004-07-31
Date
2004-07-31
Author(s)
張帆人
DOI
922213E002043
Abstract
The attitude determination is very
important in the moving vehicles, the topic of
this thesis is the attitude determination from the
information that GPS and gyro provide. The unit
quaternion which expressed the attitude can be
derived from the angular velocity that gyro
measures; also we can make use of the QUEST
algorithm to obtain the unit quaternion from
double difference of GPS carrier phase. These
two data can be combined into a more accurate
one by filters.
Since the state equations are nonlinear, the
linearization approaches may be used. The
extended Kalman filter is an appropriate method
for attitude determination. But unit norm
constraint of quaternion vector results in the
singularity of the error covariance matrix. To
overcome this problem, we can use the modified
extended Kalman filter. The modified extended
Kalman filter adopts quaternion multiplication
which preserves the unit norm of quaternion and
reduces the dimension of error covariance matrix
by an assumption of infinitesimal rotation.
In addition to the modified extended
Kalman filter, we propose the concept of
spherical linear interpolation (SLERP) in the
filter. The advantage of SLERP is less
computation load and its performance is similar
to that of Kalman filter. But the method of
SLERP can not provide the measurement update
of gyro bias, so using this method is restricted by
mission duration. Therefore, when selecting the
filter, we need consider a tradeoff among
mission duration, computation load and gyro
bias.
Subjects
EKF
SLERP
Integration
Publisher
臺北市：國立臺灣大學電機工程學系暨研究所
Type
report
File(s)![Thumbnail Image]()
Loading...
Name
922213E002043.pdf
Size
87.11 KB
Format
Adobe PDF
Checksum
(MD5):190c6f9c58af914ddca76e30d5ced08f