Comparative analysis of stochastic identification methods and fault diagnosis for multi-mode discrete systems

Abstract

This paper offers the comparative analysis of stochastic approximation methods applied for multi-mode discrete systems identification from incomplete noisy observations. The main item discussed here concerns the rate of convergence of these methods and several aspects that are considered to affect this property. To corroborate the theoretical arguments, the experimental results are provided. © Springer-Verlag Berlin Heidelberg 2003.