Robust Flight Controllers.
Abstract
This study examines the concept of robustifying a controlled system against differences which may exist between the real world system and the low-order design model upon which the controller design is based. The types of controllers considered are based upon the Linear system model, Quadratic cost, and Gaussian noise process methodology of optimal control theory. It is assumed that full-state feedback is not available and a Kalman filter is employed to provide state estimates to the controller. Both continuous-time and sampled-data controllers are considered. Two robustification techniques are considered. The first is the method of injecting zero-mean white Gaussian noise into the design model at the point of entry of the control inputs during the process of tuning the Kalman filter. The second method is an extension of the first, where the white noise is replaced by a time-correlated noise. This allows the primary strength of the noise to be concentrated only in the frequency range where robustification is desired. Comparing the results of applying the two methods allows a designer to make a trade-off between the amount of desired robustification and the performance degradation at the design conditions which occurs when the techniques are applied. Both methods are found to improve substantially the robustness properties of the controllers considered.
Document Details
- Document Type
- Technical Report
- Publication Date
- Dec 01, 1983
- Accession Number
- ADA138425
Entities
People
- J. M. Howey
Organizations
- Air Force Institute of Technology