동적시스템 모형 및 제어에 있어서 Koopman 연산자 소개

Introduction to Koopman Operator in Modeling and Control of Dynamic Systems
  • 김진성
  • 정정주
Citations

SCOPUS

2

초록

This paper briefly introduces the Koopman operator framework in the modeling and control of dynamic systems. The paper reviews the theoretical foundations of the Koopman operator, presenting implications for modeling and control in engineering systems. We describe the extended dynamic mode decomposition method to approximate the Koopman operator in a finite-dimensional space. We then show how an autoencoder is obtained for the approximated Koopman operator and analyze the uncertainty quantification. Numerical simulation reveals the validity of the proposed method. We also briefly review the interdisciplinary significance of the physics-informed Koopman operator and its potential to revolutionize the analysis and control of complex dynamic systems across various domains.

키워드

Kooperman operatorModeling and controlPhysics-Informed Koopman operatorAutoencoderMachine LearningDeep Learning.Deep learningDynamic mode decompositionDynamicsLearning systemsNumerical methodsUncertainty analysis
제목
동적시스템 모형 및 제어에 있어서 Koopman 연산자 소개
제목 (타언어)
Introduction to Koopman Operator in Modeling and Control of Dynamic Systems
저자
김진성정정주
DOI
10.5302/J.ICROS.2024.24.0054
발행일
2024-04
저널명
제어.로봇.시스템학회 논문지
30
4
페이지
373 ~ 382