Design of Downlink Hybrid NOMA Transmission

Zhiguo Ding, Robert Schober, H. Vincent Poor

TL;DR

This paper proposes a "hybrid" approach to Non-Orthogonal Multiple Access (NOMA), treating it as a seamless add-on to legacy TDMA and SDMA networks rather than a disruptive replacement. The authors prove that in downlink scenarios, this hybrid scheme consistently outperforms traditional orthogonal access in energy efficiency, particularly in the emerging "near-field" communication environments expected in 6G.

What Problem Does This Paper Address?

In the quest for higher spectral efficiency in 5G and 6G, NOMA has been a hot topic. However, most research treats NOMA and Orthogonal Multiple Access (OMA)—like Time Division Multiple Access (TDMA)—as competing philosophies. This creates a deployment hurdle: how do we integrate NOMA into existing legacy networks without a total infrastructure overhaul?

Furthermore, while much is known about uplink hybrid NOMA (where users send data to a base station), the downlink (base station to users) properties are less understood. The authors set out to design a downlink hybrid NOMA framework that acts as a "plug-and-play" enhancement for legacy networks, specifically addressing energy minimization and the complexities of near-field beamforming.

Methods

The paper tackles the problem in two stages:

  1. SISO (Single-Input Single-Output) Scenario: The authors introduce a scheme where users assigned to later time slots in a TDMA schedule are encouraged to "borrow" the time slots of earlier users. They use Successive Interference Cancellation (SIC) with a specific descending decoding order to ensure that legacy users aren't disrupted. They formulate a multi-objective energy minimization problem and solve it using a successive resource allocation strategy.
  2. MISO (Multiple-Input Multiple-Output) Scenario: The model is extended to a base station with multiple antennas serving two groups of users. Here, they introduce beam sharing. Users in a far-field group (G2) utilize the spatial beams already preconfigured for a near-field group (G1). Because beam sharing introduces non-convex interference terms, the authors employ Successive Convex Approximation (SCA) to find a sub-optimal but highly efficient power allocation solution.

Datasets

As a theoretical communication theory paper, no external real-world datasets were used. Instead, the authors validated their models through extensive Monte Carlo simulations.

  • SISO simulations: Used independent and identically distributed (i.i.d.) complex Gaussian random variables for channel gains.
  • MISO simulations: Modeled a 28 GHz carrier frequency with a 512-element Uniform Linear Array (ULA). Near-field users were placed 10–50m from the base station, while far-field users were at 200m.

Key Insights and Ideas

  • NOMA as an Add-on: The "aha moment" is the realization that NOMA doesn't have to replace TDMA. By allowing a user to transmit across multiple time slots (their own plus others'), the system gains a degree of freedom that OMA lacks.
  • Downlink vs. Uplink Divergence: A surprising analytical finding is that for ordered channel gains and equal slot durations, downlink hybrid NOMA always outperforms TDMA. This contradicts previous findings in uplink studies, where OMA can sometimes be superior depending on task deadlines.
  • The Interference Invariant: In the optimal SISO solution, the accumulated interference across different NOMA time slots is identical. This mathematical symmetry simplifies the search for optimal power coefficients.
  • Solving the OMA Singularity: In MISO systems with beamforming, OMA can hit a "singularity" where the required energy to meet a target data rate becomes infinite due to inter-beam interference. Hybrid NOMA avoids this by spreading the data requirements across more time slots.

Key Results

  • Energy Efficiency: Simulations show that hybrid NOMA significantly reduces total energy consumption compared to OMA, especially as the target data rate ($R$) increases.
  • Analytical Match: For the SISO case, the derived closed-form expressions for power allocation perfectly matched the results of exhaustive numerical searches.
  • Near-Field Superiority: In MISO scenarios, the hybrid NOMA scheme using beamfocusing (tailored for near-field) outperformed standard zero-forcing beamforming, proving that NOMA is a viable way to handle the "imperfect" resolution of near-field beams.
  • Scalability: The performance gap between NOMA and OMA widens as the number of users in the legacy group ($G_1$) increases, suggesting NOMA is highly beneficial for dense 6G cells.

Limitations and Future Work

The authors acknowledge a few areas for further exploration:

  • User Grouping: The MISO model currently focuses on two groups of users. Extending this to multiple groups or more complex clustering remains a challenge.
  • Practical Channel Constraints: The study assumes perfect Channel State Information (CSI) and largely focuses on line-of-sight near-field models. Future work needs to account for non-line-of-sight (NLOS) paths and the overhead of CSI acquisition in fast-fading environments.
  • Hardware Impairments: The impact of practical beamforming hardware (like phase noise or limited resolution of phase shifters) on beam sharing was not discussed.

Summary updated

BibTeX
@article{ding2024design,
  author = {Zhiguo Ding and Robert Schober and H. Vincent Poor},
  title = {{Design of Downlink Hybrid NOMA Transmission}},
  journal = {IEEE Trans. Wireless Communications},
  year = {2024},
  doi = {10.1109/twc.2024.3485224},
}
Figure