A Computer-Oriented Study of Performance Feedback in Sampled-Data Control Systems

摘要

This research is concerned with the development of computer-oriented procedures for the design of sampled data control systems. Two general areas of this problem are considered: the synthesis of optimal controllers having low dynamical order, and the design of adaptive control systems. The performance of the system is evaluated in terms of a quantitative performance criterion. Two new performance criteria are developed; the meanquadratic-error criterion and the average rate-of-decay criterion. Since the primary requirement of a dynamical system is stability, these criteria are examined in terms of the theorems from Liapunov's direct method of stability analysis. The analysis of the rate criterion may be applied to nonlinear as well as linear systems. Included in the discussion are illustrations on how some of the qualitative performance criteria, such as 'good' transient response, may be satisfied through proper selection of the mathematical performance criterion. The optimization procedures are based upon the concept of performance feedback in which the evaluated performance of the system is used directly in the adjustment of the system parameters.