Fuzzy model validation using the local statistical approach

  • Authors:
  • G. Rigatos;Q. Zhang

  • Affiliations:
  • Unit of Industrial Automation, Industrial Systems Institute, 26504 Rion Patras, Greece;INRIA, Campus Universitaire de Beaulieu, 35402, Rennes Cedex, France

  • Venue:
  • Fuzzy Sets and Systems
  • Year:
  • 2009

Quantified Score

Hi-index 0.20

Visualization

Abstract

The local statistical approach for fault detection and isolation is applied to fuzzy models validation. The method detects the inconsistencies between a fuzzy rule base and the modelled system. It can also identify which are the faulty parameters of the fuzzy model. The Fisher information matrix explains the detectability of changes in the parameters of the fuzzy model. Simulation tests illustrate the method's credibility.