Robust Discrete Estimation of the Space Shuttle Main Engine.
Abstract
This thesis applies recently developed robust H at infinity, or game theoretic, estimation algorithms to the Space Shuttle Main Engine (SSME). The objective is to process noisy, inaccurate sensor data in order to obtain estimates of pressure in the main combustion chamber and the oxygen to fuel mixture ratio. Each of the estimators are based on discrete time, state space models of the SSME, and employ varying levels of robustness when solving the H at infinity estimation problem. Two general problems are examined. First, H at infinity minimax estimators are derived for the case where the plant dynamics are accurately known, but the noise statistics are uncertain. The effects of various noise inputs are explored. Next, robust H at infinity estimators are designed when plant, sensor, and noise uncertainties are present. It is shown that the performance of the normally optimal Kalman filter degrades considerably in the presence of model uncertainty. By contrast, the robust H at infinity estimators perform well for the entire range of plant, sensor, and noise models considered.
Document Details
- Document Type
- Technical Report
- Publication Date
- Jun 01, 1996
- Accession Number
- ADA311712
Entities
People
- Jonathan A. Jensen
Organizations
- Air Force Institute of Technology