Introduction
As the industry moves toward 5G and beyond, the demand for higher bandwidth and improved spectral efficiency has never been greater. Millimeter-wave (mmWave) frequencies are a primary candidate for meeting these needs due to their massive available bandwidth. However, mmWave signals face significant challenges, including severe path loss and high sensitivity to physical blockages. Non-Orthogonal Multiple Access (NOMA) has emerged as a potential solution to enhance spectral efficiency by allowing multiple users to share the same frequency and time resources through power-domain multiplexing.
Existing research often simplifies mmWave networks as noise-limited or assumes uniform user distributions. In reality, users tend to cluster around specific locations. This paper addresses this gap by modeling a clustered mmWave network with realistic interference and directional beamforming to determine if NOMA maintains its performance advantage over traditional Orthogonal Multiple Access (OMA) under real-world conditions.
System Model and Methodology
The researchers employ a Poisson Cluster Process (PCP), specifically a variant of the Thomas cluster process, to simulate the network. In this model, Base Stations (BSs) are located at the center of clusters, with users distributed around them following a symmetric normal distribution.
To implement NOMA, a "Nearest-Random" user pairing scheme is used: the BS pairs the user closest to it (the "near user") with another user selected at random from the cluster (the "far user"). The study incorporates a realistic Uniform Linear Array (ULA) antenna model and a Line-of-Sight (LOS) disc blockage model. Performance is evaluated through derived closed-form expressions for coverage probability and system rate, which were validated using Monte Carlo simulations based on standard 5G mmWave specifications (e.g., 28 GHz carrier frequency).
Key Insights: The Antenna Paradox and Clustering
A significant finding of this study is the relationship between antenna configuration and user performance. While increasing the number of antenna elements ($M$) generally improves the coverage for the "near user" by narrowing the beam, it can negatively impact the "far user." Because the far user is randomly located, an excessively narrow beam is more likely to miss them. This creates an "antenna paradox" where more antennas are not always better for the system as a whole; for the far user, a smaller number of antennas (e.g., $M=2$) is often better to ensure they stay within the beam's footprint.
Furthermore, the variance of user positions within a cluster ($\sigma^2$) plays a critical role. As users become more spread out (higher variance), the performance gains associated with the clustered model begin to diminish, highlighting the importance of accurate spatial modeling in network planning.
Comparative Results: NOMA vs. OMA
The analysis demonstrates that NOMA consistently outperforms OMA in terms of total system rate, particularly when power allocation is optimized.
As shown in the results, NOMA with a low power allocation to the near user ($a_k = 0.1$) performs best in high-SNR regions. The adjustable power allocation coefficient ($a_k$) is a key advantage of NOMA, allowing it to outperform OMA across a wide range of Signal-to-Noise Ratios (SNR). The mathematical models developed in this paper showed negligible differences from simulation results, confirming their reliability for network planning.
Conclusion and Future Directions
This research confirms that NOMA is a superior choice for clustered mmWave networks compared to OMA, provided that antenna configurations and power allocation are optimized. However, several areas remain for future exploration. The study assumes perfect Successive Interference Cancellation (SIC); in practice, residual interference could degrade NOMA's advantage. Additionally, the model assumes static users, whereas real-world mobility would require frequent beam realignment. Future work could also investigate alternative user pairing strategies beyond the "Nearest-Random" scheme.
Summary updated