Automated steering : fault detection and the effect of noise /
Simulation studies were conducted investigating the application of fault detection and identification (fdi) techniques as a pre-diagnostic before allowing vehicles to enter an automated highway. This application is unique in that the fault is present at the onset of the test and that the test may on...
| Main Author: | |
|---|---|
| Format: | Thesis Book |
| Language: | English |
| Published: |
[Place of publication not identified] :
[publisher not identified] ;
1999.
|
| Subjects: | |
| Online Access: | http://proxy.library.tamu.edu/login?url=http://proquest.umi.com/pqdweb?did=730318901&sid=1&Fmt=2&clientId=2945&RQT=309&VName=PQD |
| Summary: | Simulation studies were conducted investigating the application of fault detection and identification (fdi) techniques as a pre-diagnostic before allowing vehicles to enter an automated highway. This application is unique in that the fault is present at the onset of the test and that the test may only last for a short time, current fdi applications monitor vehicles the entire time they operate on the automated highway. Other difficulties associated with this application include widely varying plant parameters and noisy environment. The performance of two model-based fdi methods, innovations and fault detection filter, were evaluated and compared. The innovations fdi consistently detected the fault faster than the fault detection alter fdi and was less sensitive to changes in vehicle parameters and operating conditions. However the fault detection filter fdi was affected less by non-Gaussian noise and was far superior to the innovations fdi in its ability to correctly differentiate between results due to small faults and to fault indications caused by noise. The trade-off between increased length of testing time and false alarm rate was also studied. For both fdi methods it was shown that the standard deviation of the false alarm rate dropped sharply to a steady state values as the test length was increased. In related work, the propagation of sensor noise in a state estimation scheme was studied. A lower bound for the steady state error covariance magnitude is derived as a function of the sensor noise covariance and system matrices. This provides design considerations for minimizing the effect sensor noise has on state estimation error. |
|---|---|
| Item Description: | Vita. "Major Subject: Electrical Engineering". |
| Physical Description: | x, 112 leaves : illustrations ; 28 cm. Issued also on microfiche from University Microfilm Inc. |
| Bibliography: | Includes bibliographical references (leaves 107-109). |