Joint Optimization of Uplink and Downlink Power in Full-Duplex Integrated Access and Backhaul
Giovanni Interdonato, Silvia Mura, Marouan Mizmizi, Stefano Buzzi, Umberto Spagnolini
TL;DR
This work targets joint uplink-downlink power optimization in in-band full-duplex mmWave IAB networks acting as range extend ers for beyond-5G. It proposes a centralized optimization framework with two PA objectives—max-min fairness and max-sum $SE$—and stabilizes the problem via MRC and SVD-based precoding to reformulate it as a geometric program with global optimality. Numerical results show that power optimization significantly improves spectral efficiency, especially as the number of UEs served by the IAB node grows, with max-sum $SE$ generally outperforming max-min fairness, though the latter can be advantageous in low IAB UE scenarios. A key finding is that the backhaul link tends to be rank-1 due to LOS geometry, imposing a fundamental bottleneck on backhaul capacity and spatial multiplexing, guiding future exploration toward multi-hop and higher-rank backhaul designs.
Abstract
We examine the performance of an Integrated Access and Backhaul (IAB) node as a range extender for beyond-5G networks, focusing on the significant challenges of effective power allocation and beamforming strategies, which are vital for maximizing users' spectral efficiency (SE). We present both max-sum SE and max-min fairness power allocation strategies, to assess their effects on system performance. The results underscore the necessity of power optimization, particularly as the number of users served by the IAB node increases, demonstrating how efficient power allocation enhances service quality in high-load scenarios. The results also show that the typical line-of-sight link between the IAB donor and the IAB node has rank one, posing a limitation on the effective SEs that the IAB node can support.