System Identification for Robust Design

本研究的目的是研究稳健的系统辨识方法。鲁棒系统识别算法产生适合于鲁棒设计的模型。特别是，所得到的模型伴随着识别误差的量化，这使得后续的设计可以针对一系列可能的系统进行。该研究项目所设想的稳健系统识别方法与来自控制和通信理论的稳健设计程序相结合，将使重要技术能够以比目前更高的性能和可靠性运行。本计划将解决的研究问题是现有的鲁棒系统识别方法的持续发展（特别是解决实际实施问题），将鲁棒系统识别方法扩展到以前未解决的系统模型类（特别是时变和有限时间持续过程），以及将系统识别算法开发到重要技术领域的应用。系统识别算法的应用发展到重要的技术领域。相关的应用研究是拟议研究计划的一部分，涉及水下声学的鲁棒检测技术和柔性结构的高性能控制。＊＊＊

9303356 Jacobson The purpose of the proposed research is to study robust system identification methods. Robust system identification algorithms produce models that are appropriate for use in robust design. In particular, the resulting models are accompanied by a quantification of the identification error which allows a subsequent design to proceed against a family of possible systems. The combination of the envisioned robust system identification methods from this research program with robust design procedures from control and communication theory will enable important technologies to operate at higher performance and reliability than they currently are able to attain. The research issues that will be addressed in this program are the continued development of existing robust system identification methods (in particular, to address practical implementation issues), the extension of robust system identification methods to classes of system models not previously addressed (in particular, to time varying and finite time duration processes) and the application of the system identification algorithms developed to important technological areas. The application of the system identification algorithms developed to important technological areas. The relevant application studies that are part of the proposed research program are robust detection techniques involving underwater acoustics and high performance control of a flexible structure. ***