User Association and Resource Allocation in Unified NOMA Enabled Heterogeneous Ultra Dense Networks

Zhijin Qin, Xinwei Yue, Yuanwei Liu, Zhiguo Ding, Arumugam NallanathanIEEE Communications Magazine 2018 · OpenReview

Introduction

As the world moves toward billions of connected IoT devices, 5G networks face critical challenges in massive connectivity and spectrum scarcity. Heterogeneous Ultra Dense Networks (HUDNs)—which deploy small cells like pico, femto, and relays closer to users—offer a solution for density. However, traditional Orthogonal Multiple Access (OMA) limits capacity by assigning one user per resource slice. This paper proposes a unified Non-Orthogonal Multiple Access (NOMA) framework to break these limits, enabling multiple users to share the same frequency/time resource block (RB) simultaneously.

System architecture of a NOMA-enabled HUDN showing Macro, Pico, Femto BSs, and Relays.

The Unified NOMA Framework

The core innovation is a unified architecture that integrates both power-domain (PD-NOMA) and code-domain (CD-NOMA) variants into a single sparse-matrix model. In this matrix, a "1" represents a user occupying a specific RB, while a "0" indicates no occupancy.

  • PD-NOMA: Users are multiplexed on a single carrier (one row of the matrix) using different power levels. Successive Interference Cancellation (SIC) is used at the receiver to separate signals.
  • CD-NOMA: This multi-carrier extension maps data streams across multiple RBs using sparse patterns or low-density spreading codes, typically decoded via the Message Passing Algorithm (MPA).
Uplink and Downlink NOMA models illustrating the sparse matrix mapping of users to resource blocks.

This unified approach allows the network to adapt: PD-NOMA is ideal for mobile users with varying channel conditions, while CD-NOMA provides the shaping gains necessary for high-reliability scenarios.

Network Management and Resource Allocation

Managing an ultra-dense network requires sophisticated strategies for user association and resource distribution to mitigate interference.

Flexible User Association

Instead of simply connecting to the nearest base station (BS), users in this framework associate based on the maximum average received power, accounting for NOMA-specific power-sharing coefficients. As network density increases, this strategy effectively offloads traffic from macro cells to small cells. For instance, once pico cell density exceeds a certain threshold, macro cell association probability drops significantly, proving the framework's ability to manage edge traffic.

Intelligent Resource Allocation

The authors employ matching theory and sequential convex programming to solve the NP-hard problem of assigning users to RBs and optimizing power levels. This dual approach minimizes both intra-cell and inter-cell interference, ensuring that the benefits of NOMA are not negated by the noise of a dense deployment.

Key Results and Performance

Simulations using stochastic geometry (modeled via Poisson point processes) demonstrate that NOMA-enabled HUDNs consistently outperform OMA in several key metrics:

  • Sum Rate Improvement: NOMA achieves a significantly higher total data rate (bits/s/Hz) compared to OMA as the number of base stations increases.
  • Fairness: Using Jain’s fairness index, the study shows that NOMA provides a more equitable distribution of resources, even under high network stress.
  • Diversity Gain: By allowing more users to multiplex on the same RB, the system captures higher multi-user diversity, further boosting the overall sum rate.

Conclusion and Future Directions

While the unified NOMA framework offers clear advantages in spectral efficiency and connectivity, challenges remain. The computational complexity of the Message Passing Algorithm (MPA) for CD-NOMA is a hurdle for standardization. Additionally, the trade-off between NOMA's high data rates and the battery life of IoT devices requires further investigation. Future work will likely focus on integrating Machine Learning to adaptively switch between NOMA modes and developing Software Defined Radio (SDR) testbeds for real-world validation.

Summary updated

BibTeX
@article{qin2018user,
  author = {Zhijin Qin and Xinwei Yue and Yuanwei Liu and Zhiguo Ding and Arumugam Nallanathan},
  title = {{User Association and Resource Allocation in Unified NOMA Enabled Heterogeneous Ultra Dense Networks}},
  journal = {IEEE Communications Magazine},
  year = {2018},
  url = {https://openreview.net/forum?id=ke4xuTnwEY},
}
Figure