A PF-Enhanced Hybrid CQI–Interference Scheduler for MU-MIMO with Long-Term Clustering and RZF Precoding

Abstract

An essential part of 5G networks are multi-user multiple-input multiple-output (MU-MIMO) systems. MU-MIMO systems will make it possible to use the spectrum more effectively, enhance user experiences by increasing throughput, and enable a higher level of throughput under dense deployment. However, conventional scheduling techniques like hybrid channel-aware prioritization (HCAP) mainly rely on current channel quality measurements, which can result in less-than-ideal resource utilization, lower levels of fairness, and starvation problems for certain users, especially when it comes to heterogeneous and changing interference scenarios. The proposed architecture, which is outlined in this paper, presents a fairness-optimized hybrid adaptive scheduler that combines HCAP and PF weighting with both a clustering approach to grouping users and a quota-based resource allocation strategy to result in a better throughput-to-fairness balance. Additionally, the proposed framework employs a merging strategy known as the "tiny-cluster" merger, as well as a regularized zero-force (RZF) precoding method to improve the ability to employ spatial multiplexing and mitigate interference. Results from simulation studies run on MATLAB described in this paper for 5G downlink scenarios with multiple users demonstrate that this proposed scheduling approach achieves approximately 99% fairness (Jain Index ≈ 0.99), eliminates user starvation, and increases the overall throughput of the system by between 15 and 25 percent when compared to existing HCAP-only and PF-only scheduling approaches. As such, the proposed scheduling scheme represents a scalable and efficient means for the next generation of 5G MU-MIMO schedulers.