碩士 / 國立成功大學 / 航空太空工程學系 / 83 / The adaptive tuning approach has been a successful tool to
recover the system divergence of Extend Kalman filter (EKF) due
to the poor dynamic modeling. This study uses the fuzzy logic
method to provide the adaptive tune in the covariance matrix of
process noise. The approach is applied to the ground navigation
problem. In this study, one single parameter in the covariance
matrix is given in two ways: 1) a constant specified value; 2)
a fuzzy tuned value according to the designed fuzzy logic
rules. The first way is referred to as "Constant-Q" approach
which is used as a comparison with the fuzzy tuned results and
a checking point for appropriate parameter values. The later
way is re- ferred to as "Fuzzy Logic Tuned (FLT)" approach. To
assess the approaches, two experiments of car navigation have
been con- ducted. The first experiment is to assess the
capability of FLT in position estimation. The second one tests
the capability of FLT under the external noise. The results
show that the designed FLT-EKF method is capable of providing
ground navigation even under a dynamic mismodeling situation.
In addition, it can successfully tolerate serious ground signal
reflection.
| Identifer | oai:union.ndltd.org:TW/083NCKU0295054 |
| Date | January 1995 |
| Creators | J.J. Chen, 陳志政 |
| Contributors | Ho, Ching-Shun, 何慶雄 |
| Source Sets | National Digital Library of Theses and Dissertations in Taiwan |
| Language | zh-TW |
| Detected Language | English |
| Type | 學位論文 ; thesis |
| Format | 48 |
Page generated in 0.0054 seconds