Abstract
This paper compares the sum rates and rate regions achieved by power-domain NOMA (non-orthogonal multiple access) and standard massive MIMO (multiple-input multiple-output) techniques. We prove analytically that massive MIMO always outperforms NOMA in i.i.d. Rayleigh fading channels, if a sufficient number of antennas are used at the base stations. The simulation results show that the crossing point occurs already when having 20-30 antennas, which is far less than what is considered for the next generation cellular networks.
Introduction
As wireless networks transition toward 5G and 6G, the goal is to serve more users with higher data rates using limited radio spectrum. Two main contenders for achieving this are Non-Orthogonal Multiple Access (NOMA) and Massive MIMO (mMIMO). NOMA serves multiple users on the same frequency/time resource by assigning different power levels and using Successive Interference Cancellation (SIC) at the receiver. Massive MIMO, on the other hand, uses a very large number of antennas at the base station to create narrow, targeted beams for each user through spatial multiplexing.
This research provides a rigorous theoretical and numerical comparison between these technologies, challenging the notion that NOMA is a necessary evolution for multi-antenna systems. The authors prove that standard Massive MIMO naturally outperforms NOMA in typical fading environments once a base station reaches a modest number of antennas.
Methodology
The authors utilize a mathematical framework to compare the ergodic sum rates of both systems in a single-cell setup with $M$ antennas and $K$ users.
- NOMA Setup: Users are paired (typically one near the center and one at the cell edge). Zero-Forcing (ZF) beamforming is used to prevent groups from interfering with each other, while SIC manages interference within each pair.
- mMIMO Setup: Standard ZF beamforming is applied to all users simultaneously, providing each user with their own spatial "lane."
To ensure a fair comparison, the authors intentionally used assumptions that favored NOMA, such as perfect channel knowledge for NOMA users and upper-bound rate expressions, while using conservative lower-bound expressions for mMIMO.
Key Insights
- Array Gain Advantage: In Massive MIMO, every user benefits from "array gain"—the ability of many antennas to focus energy. In NOMA, the weak cell-edge user often misses this gain because beamforming is optimized primarily for the strong user in the pair.
- Mathematical Suboptimality: The authors prove that if the goal is to maximize the total cell data rate, the optimal strategy in a NOMA system is to turn off cell-edge users entirely, which effectively turns the system into a suboptimal version of Massive MIMO.
- Channel Hardening: mMIMO benefits from predictable channels, allowing it to skip downlink pilots and save overhead that NOMA must expend to enable SIC.
Results and Discussion
The analytical proof demonstrates that as the number of antennas ($M$) grows, Massive MIMO always outperforms NOMA. Simulation results show that the "massive" threshold is lower than expected; the crossing point where mMIMO overtakes NOMA occurs at just 20–30 antennas—well within the range of current 5G deployments.
In specific scenarios, such as the two-user case, mMIMO can become superior with as few as 9 antennas. NOMA only maintains an advantage in the "small-scale" regime where the number of antennas is nearly equal to the number of users.
Conclusion
This study concludes that Massive MIMO is the superior choice for high-capacity cellular networks in Rayleigh fading environments. While NOMA offers theoretical benefits in power-domain stacking, the spatial resolution provided by 20 or more antennas makes NOMA's complex interference cancellation techniques redundant. Future research may further explore these trade-offs in multi-cell environments and sparse Line-of-Sight channels.
Summary updated