Performance Analysis of Self-Interference Cancellation in Full-Duplex Large-Scale MIMO Systems

This paper presents the performance analysis of the self-interference cancellation (SIC) methods in full-duplex largescale multiple-input multiple-output systems. To mitigate selfinterference (SI), we assume that the full duplex-base station (BS) uses SI-subtraction or spatial suppression. Analytical and numerical results confirm that the SI-subtraction outperforms the spatial suppression for SIC in a perfect channel estimation case. It is also concluded that the uplink and overall ergodic rates performance of the spatial suppression is respectively better than those of the SI-subtraction in a imperfect channel estimation case under a given system constraint such as uplink/downlink sum rates and the total transmit power at the BS.