CAREER: Robust Design of Adaptive Nonlinear Control

9502945 Kanellakopoulos由于大多数物理系统包含非线性和未知参数，非线性系统的自适应控制直接关系到实际应用。首席调查员之前的研究产生了几个新的系统设计程序，这些程序在不增加控制工作量的情况下，在大范围作业区域内提供高性能，并且不需要进行全状态测量。所提出的研究计划将继续发展新一代自适应系统，其中控制器的非自适应部分能够保证有界性，而自适应部分仅用于改善暂态和渐近性能。将开发更多的设计工具，这些工具不仅可以产生更强的稳定性和性能，而且还可以适用于更大类的非线性系统，并通过几何条件来表征这些类。这些新工具将应用于对商业应用具有重要意义的问题，如感应电机、智能高速公路上重型车辆的自动化操作和半导体制造工艺，并将与感兴趣的行业赞助商开展合作研究工作。拟议的教育计划将开发几门新的关于线性、非线性和自适应控制的本科生和研究生课程，以及一门本科实验室控制课程。此外，它还将通过早期介绍非线性实例和概念来提高控制系统的教学效果，以激发学生的兴趣，并确保他们理解线性控制理论只能对大多数物理系统产生区域性结果。此外，教育计划将图形和符号软件纳入本科生和研究生控制课程，这样学生可以花更少的时间推导复杂的表情并将它们编码到模拟包中，而花更多的时间探索设计方案，比较不同的方法，并通过动画和视觉模拟观看他们努力的结果。***

9502945 Kanellakopoulos Since most physical systems contain nonlinearities and unknown parameters, adaptive control of nonlinear systems is directly relevant to practical applications. The Principal Investigator's previous research resulted in several new systematic design procedures which deliver high performance over large operating regions without an increase in control effort and without requiring full-state measurements. The proposed research plan will continue the development of a new generation of adaptive systems, in which the nonadaptive part of the controller is capable of guaranteeing boundedness, while adaption is used only for the improvement of transient an asymptotic performance. Additional design tools will be developed that not only result in even stronger stability and performance properties, but are also applicable to larger classes of nonlinear systems, and characterize those classes via geometric conditions. The new tools will be applied to problems which are of importance to commercial applications such as induction motors, the automated operation of heavy-duty vehicles on intelligent highways and semiconductor manufacturing processes, and will develop cooperative research efforts with interested industrial sponsors. The proposed education plan will develop several new undergraduate and graduate courses on linear, nonlinear and adaptive control, as well as an undergraduate laboratory control course. Furthermore, it will enhance the teaching effectiveness of control systems through the early introduction on nonlinear examples and concepts in order to stimulate student interest and ensure their understanding of the fact that linear control theory can only yield regional results for most physical systems. In addition, the education plan includes graphical and symbolic software into the undergraduate and graduate control curriculum, so that students can spend less time deriving complicated expressions and coding them into a simulatio n package, and more time exploring design options, comparing different approaches, and watching the results of their efforts come alive through animation and visual simulation. ***