Steady-State Performance Analysis of Nonlinear Spline Adaptive Filter Under Maximum Correntropy Criterion

Correntropy has recently attracted great attention because of its robustness to non-Gaussian noise. Inspired by this robustness property, a spline adaptive filtering algorithm based on the maximum correntropy criterion (SAF-MCC) has been proposed. However, the steady-state convergence performance of the SAF-MCC has not been studied. This brief presents a steady-state performance analysis of the SAF-MCC for Gaussian and non-Gaussian noise. We derive and discuss the theoretical excess mean square error (EMSE) under these cases. Monte Carlo simulations are presented to corroborate the theoretical performance analysis results developed in this brief.