Research Initiation Award: Parameter Estimation in Adaptive Control

研究启动奖：自适应控制中的参数估计

• 批准号: 9111346
• 负责人: Konstantinos Tsakalis
• 金额: $ 6万
• 依托单位: Arizona State University
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 1991
• 资助国家: 美国
• 起止时间: 1991-09-01 至 1994-02-28
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=9111346&HistoricalAwards=false
• 关键词: Research; Initiation; Award; Parameter; Estimation

The objective of the research is to analyze and quantify the design of estimators with intended application in adaptive control systems. Interest is mainly focused on certain aspects of closed-loop performance in the presence of uncertainty and, particularly, on preventing the closed loop signals from reaching excessively large magnitude for short time periods, commonly known as "burst" phenomena. Such undesirable phenomena are of great practical significance, especially in cases of poorly excited systems where the true nonlinear nature of adaptive controllers is highly pronounced. The plan is to investigate the properties of estimators whose design is performed via signal energy as well as amplitude criteria. Although the minimization of the maximum amplitude of the estimation error typically leads to a linear programming problem of increasing dimensionality, recursive suboptimal solutions will be considered in order to alleviate this problem. The preliminary investigations indicate that these techniques allow for amplitude considerations in the design of adaptive systems and for a more efficient use of the available measurements. Particular emphasis will be given to the extension of this approach to time-varying systems and the assessment of the associated performance trade-offs. It is expected that the results of the research will make a significant contribution in the design of practically acceptable adaptive controllers.

研究的目的是分析和量化在自适应控制系统中预期应用的估计器的设计。兴趣主要集中在存在不确定性时闭环性能的某些方面，特别是防止闭环信号在短时间内达到过大的幅度，通常称为“突发”现象。这种不良现象具有重要的实际意义，特别是在弱激励系统中，自适应控制器的真正非线性性质非常明显。该计划是研究估计器的性质，其设计是通过信号能量和幅度标准来执行的。虽然估计误差最大幅度的最小化通常会导致一个增加维数的线性规划问题，但为了缓解这个问题，将考虑递归次优解。初步研究表明，这些技术允许在设计自适应系统时考虑幅度，并更有效地利用现有的测量值。将特别强调将这一方法扩展到时变系统和评估有关的性能权衡。期望研究结果对设计实际可接受的自适应控制器有重大贡献。