NSF Young Investigator

点击翻译按钮获取中文摘要

9358288 Rotea As in other areas of engineering, design of control systems involves tradeoffs among competing objectives. The basic problem is to find a controller that optimizes several competing performance and robustness requirements associated with one dynamical system. Good tracking of reference inputs vs low sensitivity to sensor noise, robust stability vs performance, are just a few examples of these tradeoffs. Clearly, it is important to develop tools to help the designer understand how the various competing objectives conflict with each other. One goal of this research is to obtain design procedures for the synthesis of controllers that meet several design specifications in the presence of uncertainty in the plant description. It is intuitively obvious that the performance of a given system may be greatly improved if the plant to be controlled and the controller are designed simultaneously rather than designing the plant first and then the controller. Unfortunately, in some cases of engineering relevance, an integrated plant/controller design is not feasible due to the large number of variables and design objectives. However, there are some challenging aerospace problems for which such an integration is not only feasible but also desirable. Another goal of this research is to develop a methodology for the design of systems (i.e. both plant and controller are to be designed) that satisfy multiple design specifications of practical significance. This research has three main components: (1) the development of new theoretical results for the analysis and synthesis of robust feedback systems with multiple design specifications, (2) the development of computational efficient algorithms to implement the theory, and (3) applications to validate the results. ***