Polynomial Series FLANN for Nonlinear Equalization

Efficient equalization for nonlinear communication channels with Additive White Gaussian Noise (AWGN) is presented. The proposed equalization is based on a Functional Link Artificial Neural Network (FLANN) structure in which the original input is nonlinearly expanded. The proposed nonlinear expansion follows a polynomial series. The nonlinearity incorporated at the output of the conventional FLANN is omitted in the proposed Polynomial Series Equalizer (PSE). Consequently, the convergence of the PSE is fast and its computational complexity is low. Moreover, explicit mathematical formula for the optimum PSE is obtained. The PSE is adapted using the fast gradient based signed Least Mean Squared (LMS). Simulations demonstrate that, the PSE vastly outperforms other FLANN based equalizers employing the Bit Error Rate (BER) metric at different nonlinear channel models and different Signal to Noise Ratios (SNR).