Reference Receiver Based Digital Self-Interference Cancellation in MIMO Full-Duplex Transceivers

TL;DR

This paper introduces a novel reference receiver-based digital self-interference cancellation method for MIMO full-duplex transceivers, significantly improving interference suppression with low-complexity linear digital processing.

Contribution

It proposes a new cancellation structure using reference receiver chains, not previously discussed for full-duplex transceivers, and analyzes its performance and potential bottlenecks.

Findings

Achieves better self-interference suppression than existing solutions.

Effective under practical RF/analog impairment levels.

Uses only low-complexity linear digital processing.

Abstract

In this paper we propose and analyze a novel self-interference cancellation structure for in-band MIMO full-duplex transceivers. The proposed structure utilizes reference receiver chains to obtain reference signals for digital self-interference cancellation, which means that all the transmitter-induced nonidealities will be included in the digital cancellation signal. To the best of our knowledge, this type of a structure has not been discussed before in the context of full-duplex transceivers. First, we will analyze the overall achievable performance of the proposed cancellation scheme, while also providing some insight into the possible bottlenecks. We also provide a detailed formulation of the actual cancellation procedure, and perform an analysis into the effect of the received signal of interest on self-interference coupling channel estimation. The achieved performance of the proposed reference receiver based digital cancellation procedure is then assessed and verified with full waveform simulations. The analysis and waveform simulation results show that under practical transmitter RF/analog impairment levels, the proposed reference receiver based cancellation architecture can provide substantially better self-interference suppression than any existing solution, despite deploying only low-complexity linear digital processing.