Robust Identification of Neuro-Fractional-Order Hammerstein Systems

This paper introduces a neuro-fractional order Hammerstein model with a systematic identification algorithm, which is robust against non-Gaussian measurement noises and outliers. The proposed model consists of a Radial Basis Function (RBF) in series with a Fractional-Order System (FOS). The proposed identification scheme is accomplished in two stages. The fractional order of the FOS is estimated in the frequency-domain. Then, the weights of the RBF and the coefficients of the FOS are determined in the time domain via Lyapunov stability theory. Real measurement data contain outlier, which badly degrades the results of conventional identification algorithms. To overcome this difficulty a correntropy kernel-based Lyapunov function is proposed that is robust against outliers. The effectiveness of the proposed method is illustrated through a simulating example.