Earticle

초록

영어

The class of Basis Function Piecewise-affine (BPWA) functions is an ideal model structure for nonlinear and hybrid system identification. This paper proposes a parameter space decomposition theorem for BPWA functions. It is proved that any BPWA function has a decomposed parametric representation, in which the algebraic and geometrical parameters can be identified separately. Under the assumption of invariant geometrical structure, a recursive algorithm is proposed to identify the BWPA AutoRegressive Exogenous models from the input-output data. Two benchmark examples are illustrated to show that the proposed algorithm has much higher computational efficiency compared with the competing algorithms with the same identification accuracy.

목차

Abstract
 1. Introduction
 2. BPWA Functions
  2.1 BPWA Basis Function
  2.2 BPWA Approximation Theorem
 3. Parameter Space Decomposition Theory of BPWA Functions
 4. Recursive Identification of PWA Systems
  4.1 BPWARX Model [17]
  4.2 Change of Regression Vector
  4.3 Recursive Identification Algorithm
 5. Numerical Examples
 5. Conclusions
 References

키워드

저자

• Rui Huang [ System Department, United Technologies Research Center, East Hartford, CT, USA ]
• Weiquan Sun [ School of Business Administration, Duquesne University, Pittsburgh, PA, USA ]
• Min Li [ Swanson School of Engineering, University of Pittsburgh, Pittsburgh, PA, USA ]

참고문헌

발행기관

간행물

표지보기 관심저널 등록

• 간행물명

  International Journal of Control and Automation

• 간기

  월간

• pISSN

  2005-4297

• 수록기간

  2008~2016

• 십진분류

  KDC 505 DDC 605