Practical Methods for the Compensation and Control of Multivariable Systems.
BROWN UNIV PROVIDENCE R I DEPT OF ENGINEERING
Pagination or Media Count:
A new parameter adaptive control scheme for linear multivariable systems has been developed. It is felt that such an adaptive controller could be used in a variety of aerospace applications. A multi-purpose controller has been designed which simultaneously decouples, places poles arbitrairily, rejects disturbances, insures zero error tracking, and is robust with respect to parameter variations. A new and straightforward method for obtaining simple low order models of systems whose dynamical behavior approximates that of more complex, higher order systems has also been developed. Such low order models can be used in the design of low compensators for the more complex systems. A complete new resolution has been presented to the question of what changes occur to the individual transfer matrix of elements of a linear multivariable system under local, scalar output feedback. In particular, it has been shown what poles become controllable and observable via any inputoutput pair when constant gain output feedback is applied between any i-th output and any j-th input. The question of parameter variation and the development of compensators which are insensitive to that variation has been resolved for a specific feedback group and is being studied for systems with parameters. Results have also been obtained for the pole-assignment problem involving parameters using intersection theory and some preliminary work has been done on the realization, coprime factorization and trace assignment problems for systems defined over the polynominal ring in n-variable over the integers. Author
- Statistics and Probability