This paper proposes a power allocation framework based on the per-user quality-of-experience (QoE) conditions for an aerial relay massive MIMO (mMIMO) network, assuming the direct transmission between the mMIMO base station and ground users (UEs) is unavailable. We first derive closed-form spectral efficiency expressions for the mMIMO-based system, with a UAV acting as the aerial relay. Then, we formulate a joint optimization problem of power allocation and QoE in the downlink, aiming to maximize the sum throughput of the serving ground UEs. The problem is generally hard to solve due to the non-convex constraints and non-concave objective function. To address it, we propose a two-step algorithm based on inner convex approximation (ICA) method. However, the ICA-based algorithm requires an initial feasible point, which is difficult to find by generating a random point as commonly conceived in existing approaches while satisfying the QoE constraints. To tackle this issue, we develop a max-min problem, whose solution leads to an initial feasible point that maximizes the difference between the per-user rate and its corresponding QoE threshold. Numerical results are used to demonstrate the validity of the theoretical analysis and the effectiveness of the proposed algorithms.
QoE-Aware Power Allocation for Aerial-Relay Massive MIMO Networks
Sanguinetti L.Ultimo
2024-01-01
Abstract
This paper proposes a power allocation framework based on the per-user quality-of-experience (QoE) conditions for an aerial relay massive MIMO (mMIMO) network, assuming the direct transmission between the mMIMO base station and ground users (UEs) is unavailable. We first derive closed-form spectral efficiency expressions for the mMIMO-based system, with a UAV acting as the aerial relay. Then, we formulate a joint optimization problem of power allocation and QoE in the downlink, aiming to maximize the sum throughput of the serving ground UEs. The problem is generally hard to solve due to the non-convex constraints and non-concave objective function. To address it, we propose a two-step algorithm based on inner convex approximation (ICA) method. However, the ICA-based algorithm requires an initial feasible point, which is difficult to find by generating a random point as commonly conceived in existing approaches while satisfying the QoE constraints. To tackle this issue, we develop a max-min problem, whose solution leads to an initial feasible point that maximizes the difference between the per-user rate and its corresponding QoE threshold. Numerical results are used to demonstrate the validity of the theoretical analysis and the effectiveness of the proposed algorithms.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.