Dimensionality reduction in kernel-based identification of Wiener system by cyclostationary excitations
DOI:
https://doi.org/10.24425/ijet.2025.153581Abstract
The topic of nonparametric estimation of nonlinear characteristics in the Wiener system is examined. In this regard, the traditional kernel algorithm faces difficulties stemming from the dimensionality associated with the memory length of the dynamic block. A particular class of input sequences has been proposed, which aids in reducing dimensionality and consequently improves the convergence rate of the estimator to the true characteristics. A theoretical analysis of the suggested method is presented.