Adaptive Non-Linear Digital Self-Interference Cancellation for Full-duplex Wireless Systems Using Ito-Hermite Polynomials

Abstract

In this paper, a novel non-linear digital self- interference cancellation algorithm is proposed for full-duplex wireless systems. The proposed algorithm constructs a replica of the nonlinearly-distorted transmitted signal using Itô generalization of the Hermite (Itô-Hermite) polynomials. Because Itô-Hermite polynomials constitute orthogonal basis functions of any non-linear function when the input is a complex Gaussian random process, the proposed least mean squares (LMS) adaptive filtering based on Itô-Hermite polynomials optimally cancels non-linearly distorted self-interference signals in full-duplex systems. From numerical results, it is shown that the proposed algorithm provides noticeable gains in terms of both a fast convergence rate and the low computational complexity for the self-interference cancellation compared to the existing LMS algorithm using Volterra- series as basis functions with basis function pre- orthogonalization. © 2018 IEEE.