Digital Self-Interference Cancellation Based on Blind Source Separation and Spectral Efficiency Analysis for the Full-Duplex Communication Systems

In this paper, the fundamental problem for the full-duplex communication systems, i.e., self-interference cancellation (SIC), is investigated, and a novel digital-domain SIC method based on blind source separation is proposed. This method achieves SIC by separation and has two prominent characteristics that are: 1) it is a blind method, and no training sequence or time is needed and 2) it is unaffected by the nonlinear components of the self-interference signal. Meanwhile, the critical factors for spectral efficiency (SE) are fully analyzed herein. In the simulations, the performances of the proposed method, including SIC capacity, signal-to-interference-plus-noise ratio, SE, and SE gain, are comprehensively evaluated and compared with least-square-based SIC algorithms. These results show that the proposed algorithm can obtain considerable performances, even the best performance in some cases.