报告题目：Minimum Input Design for Direct Data-driven Analysis of Unknown Linear Systems

报告时间：2023年4月13日15:00—17:00

报告地点：数学与统计学院310会议室

报告摘要: Modern control theory has been broadly rooted in an indirect data-driven paradigm--identifying a dynamical model followed by model-based control analysis and design. However, examples have confirmed that directly controlling a system may be potentially “easier” than identifying the model, which motivates our study of direct data-driven approach. In a sharp contrast, we analyze the system property (aka property ID) by directly using the input and state feedback data of the unknown system. Via a novel concept of sufficient richness of input sectional data, we first establish the necessary and sufficient condition for the minimum input design to excite the system for property ID. Specifically, the input sectional data is sufficiently rich for any property ID if and only if it spans a linear subspace that contains a property dependent minimum linear subspace, any vector basis of which can be easily used to form the minimum excitation input. Moreover, we rigorously show that many important structural properties can be identified with the minimum input that is however unable to identify the system model. Thus, our results quantify the advantages of the direct data-driven analysis over the model-based analysis for linear systems in terms of data efficiency.

报告人简介：游科友，清华大学自动化系长聘副教授、博士生导师。2007年获中山大学统计科学学士学位。2007年8月至2012年6月在新加坡南洋理工大学电气与电子工程学院攻读博士学位和从事博士后研究。自2012年7月起任教于清华大学自动化系。曾受邀访问意大利都灵理工大学、澳大利亚墨尔本大学、香港科技大学等院校。长期从事复杂网络化系统的学习、优化与控制及其应用研究。

目前担任 Automatica, IEEE Transactions on Control of Network Systems, IEEE Transactions on Cybernetics 等国际期刊副编委(Associate Editor)。先后主持科技创新2030-新一代人工智能重大专项(青年科学家项目)、国家自然科学基金委优青项目、重点项目、重点研发计划课题等，获中国自动化学会自然科学一等奖(排1)、关肇直最佳论文奖、亚洲控制学会(ACA) Temasek Young Educator Award奖等。