Career: Nonlinear Control: New Problems for Robust and Adaptive Design

9624386 Krstic建议的研究目标是开发利用而不是抵消工程系统的非线性特征的新的控制设计。这位首席研究员以前的研究包括解决自适应非线性控制中几个长期存在的设计问题，以及一本关于不确定非线性系统控制的新专著。这些设计在许多应用领域越来越受欢迎，并包含在商业上可用的控制设计软件中。该方案解决了具有严重非线性和高度不确定模型的系统的鲁棒自适应控制的新问题。(A)将开发避免消除有用的非线性的新设计。这是近年来PI在喷气发动机稳定化应用中得以解决的重大理论和实践问题。这一问题的一般解决方案可能会在汽车系统、飞行控制、电动马达、电力系统、制造和许多其他应用领域产生技术影响。没有消除的设计将有助于解决具有实际意义的鲁棒非线性控制问题：有限的控制功率，对噪声和未建模动态的敏感性，以及分支平衡点的稳定性。(B)发展设计方法，以解决“寻峰”问题--对不确定非线性函数的峰进行调节。这样的问题出现在许多固有的非线性应用中，其中最大效率是在峰值实现的，但不确定性阻止了设计者先验地选择设置点。(C)现有的自适应非线性控制方法将被重新设计，以允许非参数不确定性并改善未建模动态下的性能。PI最近发现，自适应控制器可以收敛到不稳定控制器(同时保证全局监管)，这推动了对自适应系统的基本几何理解，并将指导他未来的稳健重新设计。该提案的目标是出于应用问题的考虑，并将为未来与联合技术公司、福特汽车公司和罗克韦尔国际公司的合作项目提供理论基础。教育计划的重点是在PI最近出版的书的基础上开发一门新的研究生课程--非线性控制设计。将努力将与工程相关的非线性控制成果引入研究生控制课程。将特别强调使非线性控制对更广泛的控制专业学生，特别是对应用感兴趣的学生具有吸引力，以提高未来工业领袖对控制的鉴赏力。本科教学计划是UMCP本科课程重大创新改革的一部分。工业环境的变化比学术环境的变化快得多，因此，该计划要求本科生接受控制教育，传授一套持久的、具有持久价值的技能。计算机动画被提出作为本科控制教学的重要工具。***

9624386 Krstic The research objective of the proposal is to develop new control designs which exploit, rather than counteract, nonlinear features of engineering systems. The Principal Investigator's previous research includes solutions to several long-standing design problems in adaptive nonlinear control and a new monograph on control of uncertain nonlinear systems. These designs are gaining popularity in a number of application areas, and are included in commercially available control design software. This proposal addresses new problems of robust and adaptive control for systems with severe nonlinearities and highly uncertain models. (A) New designs will be developed which avoid cancellation of useful nonlinearities. This is a major theoretical and practical problem which the PI was recently able to solve for application to jet engine stabilization. A general solution to this problem is likely to have a technological impact in application areas such as automotive systems, flight control, electric motors, power systems, manufacturing, and many others. Designs without cancellation will contribute to solving robust nonlinear control problems of practical importance: limited control power, sensitivity to noise and unmodeled dynamics, and stabilization of bifurcated equilibria. (B) Design methods will be developed to solve the problem of "peak seeking"-regulation to peaks of uncertain nonlinear functions. Such problems occur in many inherently nonlinear applications where maximum efficiency is achieved at the peak but the uncertainty prevents the designer from choosing the set point a priori. (C) Existing methods of adaptive nonlinear control will be redesigned to allow nonparametric uncertainties and improve performances under unmodeled dynamics. The PI's recent finding that adaptive controllers can converge to destabilizing controllers (while guaranteeing global regulation) advances the fundamental geometric understanding of adaptive systems and will guide his future robust redesigns. The objectives of the proposal are motivated by application problems, and will provide a theoretical base for future collaborative projects with United Technologies, Ford Motor Company, and Rockwell International. The focus of the education plan is on developing a new graduate course Nonlinear Control Design based on the PI's recent book. An effort will be made to introduce nonlinear control results of engineering relevance into the graduate control curriculum. A particular emphasis will be put on making nonlinear control attractive to a broader spectrum of control students, especially those with interest in applications, so as to enhance appreciation of control among future industrial leaders. The undergraduate teaching plan is a part of a major innovative reform of the undergraduate curriculum at UMCP. The dynamics of the industrial environment are much faster than those of the academic environment, consequently, the plan calls for undergraduate control education which teaches a lasting set of skills with enduring value. Computer animation is proposed as an important tool for teaching undergraduate control. ***