A method and implementation for incorporating heuristic knowledge into a state estimator through the use of a fuzzy model /
estimation performance, as compared to a Kalman
| Main Author: | |
|---|---|
| Format: | Thesis Book |
| Language: | English |
| Published: |
[Place of publication not identified] :
[publisher not identified] ;
1998.
|
| Subjects: | |
| Online Access: | http://proxy.library.tamu.edu/login?url=http://proquest.umi.com/pqdweb?did=732834441&sid=1&Fmt=2&clientId=2945&RQT=309&VName=PQD |
| Summary: | estimation performance, as compared to a Kalman falter, when non-constant, or changing, biases exist in the measurement data. The state estimation performance increase will come from the use of a fuzzy model to determine the position and velocity gains of a state estimator. A method is proposed for incorporating heuristic knowledge into a state estimator through the use of a fuzzy model. This method consists of using a fuzzy model to determine the gains of the state estimator, converting the heuristic knowledge into the fuzzy model, and then optimizing the fuzzy model with a genetic algorithm. This method is applied to the problem of state estimation of a cascaded global positioning system (Gps)/inertial reference unit (IRU) navigation system. The GPS position data contains two major sources for position bias. The first bias is due to satellite errors and the second is due to the time delay or lag from when the GPS position is calculated until it is used in the state estimator. When a change in the bias of the measurement data occurs, a state estimator will converge on the new measurement data solution. This will introduce errors into a Kalman filter's estimated state velocities, which in turn will cause a position overshoot as it converges. By using a fuzzy model to determine the gains of a state estimator, the velocity errors and their associated deficiencies can be reduced. |
|---|---|
| Item Description: | Vita. "Major Subject: Computer Science". |
| Physical Description: | xi, 167 leaves : illustrations ; 28 cm. Issued also on microfiche from University Microfilm Inc. |
| Bibliography: | Includes bibliographical references: pages 95-98. |