Date of Award

12-1992

Document Type

Thesis

Degree Name

Master of Science in Electrical Engineering

Department

Department of Electrical and Computer Engineering

First Advisor

Peter Maybeck, PhD

Abstract

This study develops and investigates the performance of a Multiple Model Adaptive Estimator (MMAE) to detect and identify control surface and sensor failures on the LAMBDA flight vehicle (a URV developed by Wright Laboratories). The MMAE uses a bank of Kalman filters that predict the aircraft response to a given input, with each filter model based on a different failure hypothesis, and then forms the residual difference between the prediction and sensor measurements for each filter. The MMAE uses these residuals to determine the probabilities of the failures that are modeled by the Kalman filters. Initially the MMAE identified all these failures within 4 seconds of onset. Various performance improvement techniques were researched and the identification time was reduced to less than 2 seconds after failure onset. This improvement was mostly due to an increase in the penalty for measurement differences, and through returning of the Kalman filters. The MMAE performance was tested at the boundaries of the LAMBDA flight envelope, with good performance found at points close to the design flight condition. The performance at points that were far from the design flight condition indicates that gain scheduling is required to provide adequate performance across the entire envelope.

AFIT Designator

AFIT-GE-ENG-92D-19

DTIC Accession Number

ADA259137

Comments

The author's Vita page is omitted.

Download

Share

COinS