Abstract
—Integrated Access and Backhaul (IAB) enables flexible 5G deployments by reusing the same radio resources for both access and backhaul, which can enable a broad range of new use cases, particularly in sub-6 GHz bands due to their favorable propagation characteristics. However, this shared-spectrum model can suffer from limited performance due to the use of half-duplex multiplexing schemes to manage intra-node interference, if not alleviated by adoption of MIMO techniques. This paper presents a comprehensive evaluation of MIMO processing strategies for IAB systems in both uplink and downlink scenarios, with the aim of guiding the design of SDR-based IAB implementations. We investigate both Single-User (SU) and MultiUser (MU)-MIMO configurations under linear and non-linear processing schemes. Our findings show that MU-MIMO significantly outperforms SU-MIMO, particularly when using linear processing, achieving gains of up to 66% with a four-antenna IAB donor node. Furthermore, we demonstrate that non-linear processing can improve spectral efficiency by more than twice that of linear methods. Importantly, these gains can be achieved even in the presence of legacy user equipment, highlighting the potential for seamless integration of non-linear techniques into IAB networks and paving the way for the next generation of IAB deployments.