Adaptive Non-Linear Digital Self-Interference Cancellation for MIMO Full-Duplex Systems

Abstract

In this thesis, a novel non-linear digital self-interference (SI) cancellation algorithm is proposed for inband full-duplex wireless systems. The key idea of the proposed algorithm adaptively constructs a replica of the nonlinearly distorted transmitted signal using Itô generalization of the Hermite (Itô-Hermite) polynomials when the channels are changed. Using the fact that Itô-Hermite polynomials are able to constitute uncorrelated basis functions of an arbitrary non-linear function when the input follows a complex Gaussian random process, it is shown that the proposed adaptive algorithm based on Itô-Hermite polynomials optimally cancel the non-linearly distorted SI signals in full-duplex systems. From the simulations, it is shown that the proposed algorithm provides noticeable gains in terms of both the fast convergence rate and the low computational complexity for the SI cancellation compared to the existing SI cancellation algorithms. In addition, a full-duplex wireless test-bed is implemented using software defined radio to verify the feasibility of full-duplex systems and the performance of the proposed SI cancellation algorithm based on Itô-Hermite polynomials. Further, the possible extensions of the proposed algorithm are discussed in the context of mmWave and massive multiple-input multiple-output systems for beyond 5G.