Rate-dependent hysteresis modeling and compensation using least squares support vector machines

Qingsong Xu, Pak Kin Wong, Yangmin Li

Research output: Chapter in book / Conference proceedingConference article published in proceeding or bookAcademic researchpeer-review

1 Citation (Scopus)


This paper is concentrated on the rate-dependent hysteresis modeling and compensation for a piezoelectric actuator. A least squares support vector machines (LS-SVM) model is proposed and trained by introducing the current input value and input variation rate as the input data set to formulate a one-to-one mapping. After demonstrating the effectiveness of the presented model, a LS-SVM inverse model based feedforward control combined with a PID feedback control is designed to compensate the hysteresis nonlinearity. Simulation results show that the hybrid scheme is superior to either of the stand-alone controllers, and the rate-dependent hysteresis is suppressed to a negligible level, which validate the effectiveness of the constructed controller. Owing to the simple procedure, the proposed modeling and control approaches are expected to be extended to other types of hysteretic systems as well.
Original languageEnglish
Title of host publicationAdvances in Neural Networks - 8th International Symposium on Neural Networks, ISNN 2011
Number of pages9
EditionPART 2
Publication statusPublished - 6 Jun 2011
Externally publishedYes
Event8th International Symposium on Neural Networks, ISNN 2011 - Guilin, China
Duration: 29 May 20111 Jun 2011

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
NumberPART 2
Volume6676 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349


Conference8th International Symposium on Neural Networks, ISNN 2011


  • hysteresis
  • least squares support vector machines (LS-SVM)
  • motion control
  • Piezoelectric actuator

ASJC Scopus subject areas

  • Theoretical Computer Science
  • Computer Science(all)

Cite this