Feedback System Design.

Abstract

Abstract : In Quantitative Feedback Synthesis, bounds on plant uncertainty and on the system performance are specified apriori. The minimum feedback is used which satisfies the latter over the range of plant uncertainty. The principal effort during the past year was in extending Quantitative Feedback Synthesis to two new highly complex, multiple-loop, single input-output plant structural classes. The significant advantage of multiple-loop over single-loop is that, in highly uncertain systems, the same performance specifications can be satisfied, with tremendously smaller sensor noise effects. A technique denoted as 'Design Perspective' was developed, which enables the designer to evaluate the trade-offs without a detailed design. Thus, early in the game and before doing a detailed design, the designer can obtain an excellent idea of which loops (sensors) to use and which to omit and the portion of the 'feedback burden' to assign to each loop. The same design techniques are applicable to highly uncertain nonlinear plants with the same structure. Several complex detailed design examples with very large uncertainty were done. Design perspective results were very close to the final design results. The designs were simulated on the digital computer and in all cases satisfied the apriori assigned performance specifications.