Modified Ito Generalization of the Hermite Polynomials for the Linearization of Radio Over Fiber Links With Increased Numerical Stability

Abstract

Nonlinearity is a main impairment for radio over fiber (RoF) link as it can severely degrade the signal quality and limit the overall system performance. Digital predistortion (DPD) is an effective technique to linearize an RoF link. In a DPD technique, polynomials are used to model a nonlinear RoF link. However, a conventional polynomial model exhibits numerical instabilities. In this article, we introduce a novel set of orthogonal polynomials based on the modified Ito generalization of the Hermite polynomials to improve the numerical stability. Compared with other orthogonal polynomials, the proposed orthogonal polynomials have two additional advantages. First, the proposed polynomials are orthogonal for an input signal with complex Gaussian distribution. Second, the proposed orthogonal polynomials can be applied to a RoF link for multiband signal transmission. An experiment is performed to evaluate the performance of the proposed approach. Two 100 MHz orthogonal frequency division multiplexing (OFDM) signals at 1.6 and 2.4 GHz are transmitted over a RoF link in which an electro-absorption modulated laser (EML) is employed to perform signal modulation. The experimental results show that a RoF link employing the proposed orthogonal polynomials has a better performance than using the conventional polynomials in terms of error vector magnitude (EVM) and spectral regrowth suppression.