Design of an adaptive fuzzy compensator with disturbance observer using the sliding-mode technique

Yu Sheng Lu, Bing Xuan Wu, Shu Fen Lien

Research output: Contribution to journalArticle

7 Citations (Scopus)

Abstract

This paper presents an adaptive fuzzy control scheme using the sliding-mode technique for a class of nonlinear SISO systems whose input channel can have a gain function of full state variables. The proposed controller is composed of three parts: a nominal controller, a disturbance observer and an adaptive fuzzy compensator. The nominal controller specifies the desired closed-loop dynamics, while the disturbance observer and the fuzzy compensator compensate for the system perturbation, including parameter uncertainties and unexpected external disturbances. In contrast to existing adaptive fuzzy control schemes that indirectly extract perturbation information from the tracking error, the proposed scheme learns directly from a switching signal equivalent to the error of disturbance compensation; this accelerates the learning process. Moreover, the added disturbance observer enhances the performance robustness of the adaptive fuzzy system for exceptional disturbances that cannot be modeled by the fuzzy logic model. Stability analysis is provided based on Lyapunov stability theory. The experimental results concerning the tracking control of a nonlinear straight-line linkage system are also presented to illustrate the effectiveness of the proposed scheme.

Original languageEnglish
Pages (from-to)5945-5968
Number of pages24
JournalInternational Journal of Innovative Computing, Information and Control
Volume8
Issue number9
Publication statusPublished - 2012 Sep

Keywords

  • Disturbance observer
  • Fuzzy control
  • Motion control
  • Sliding mode
  • Stability

ASJC Scopus subject areas

  • Computational Theory and Mathematics
  • Information Systems
  • Software
  • Theoretical Computer Science

Fingerprint Dive into the research topics of 'Design of an adaptive fuzzy compensator with disturbance observer using the sliding-mode technique'. Together they form a unique fingerprint.

  • Cite this