NOMA Aided Narrowband IoT for Machine Type Communications With User Clustering

Ali Shahini, Nirwan Ansari

TL;DR

This paper proposes a novel Non-Orthogonal Multiple Access (NOMA) scheme combined with user clustering for Narrowband IoT (NB-IoT) to address the challenge of massive connectivity. By intelligently grouping devices and allocating resources, the proposed system significantly boosts the total throughput of NB-IoT networks, outperforming traditional OMA approaches.

What Problem Does This Paper Address?

The Internet of Things (IoT) is rapidly expanding, with projections of billions of connected devices. Next-generation mobile networks, particularly 5G, need to support diverse IoT use cases, including massive Machine Type Communications (mMTC) and ultra-reliable low-latency communications (URLLC). Narrowband IoT (NB-IoT) has emerged as a promising technology for low-power, low-cost mMTC devices, offering extended coverage and energy efficiency. However, the existing Orthogonal Multiple Access (OMA) scheme in NB-IoT, where each sub-carrier is used by only one device, struggles to provide connectivity for the sheer volume of devices expected. This limitation hinders the scalability of NB-IoT for massive IoT deployments.

The paper tackles this by exploring Non-Orthogonal Multiple Access (NOMA), a technique that allows multiple users to share the same spectral resources. Specifically, it focuses on power-domain NOMA within the NB-IoT framework, aiming to enhance network capacity and spectral efficiency. A key challenge is how to effectively manage the interference and resource allocation when multiple devices share the same resources, especially considering the differing Quality of Service (QoS) requirements of mMTC and URLLC devices.

Methods

The core of the proposed solution lies in a two-pronged approach: NOMA clustering and resource allocation.

  1. NOMA Clustering: Devices are grouped into "NOMA clusters." Within each cluster, devices share the same sub-carrier. The paper introduces a novel clustering mechanism that prioritizes URLLC devices by assigning them lower ranks (meaning they are decoded first) due to their higher data rate and power requirements. This strategy aims to mitigate interference from high-power URLLC devices to low-power mMTC devices. mMTC devices are then assigned to higher ranks within the clusters. This clustering process is designed to satisfy the QoS requirements of both device types.
  2. Resource Allocation: After clustering, the paper formulates an optimization problem to maximize the total uplink sum rate of all devices. This involves optimizing both the assignment of sub-carriers to NOMA clusters and the transmission power of each device. The problem is initially formulated as a mixed-integer nonlinear programming (MINLP) problem, which is NP-hard. To overcome this complexity, the authors propose an efficient heuristic algorithm. This algorithm iteratively allocates sub-carriers to the "best" cluster (the one that maximizes throughput) and updates device rates and powers. The power allocation sub-problem is further analyzed and shown to be a convex optimization problem after variable transformations, allowing for efficient solutions.

Datasets

Not discussed in the paper. The evaluation is based on simulations.

Key Insights and Ideas

  • NOMA for NB-IoT is a viable path to massive connectivity: The paper demonstrates that NOMA, when applied to NB-IoT, can significantly increase the number of connected devices and overall network throughput, overcoming the limitations of OMA.
  • Intelligent user clustering is crucial: Simply applying NOMA without considering device types and QoS would lead to performance degradation. The proposed ranking-based clustering, prioritizing URLLC devices, is key to managing interference and meeting diverse requirements.
  • Prioritizing URLLC devices in clustering is effective: By assigning URLLC devices lower ranks, they are decoded first, minimizing their exposure to interference from other devices within the same cluster. This is a clever way to leverage NOMA's benefits while respecting stringent URLLC requirements.
  • Heuristic algorithm provides efficient sub-optimal solutions: The NP-hard nature of the optimization problem necessitates efficient algorithms. The proposed heuristic offers a practical approach to finding good solutions without the computational burden of exhaustive search.

Key Results

  • The proposed NOMA-based NB-IoT scheme achieves an approximate 28% performance gain in total throughput compared to the OMA scenario for a large number of users.
  • The Jain's fairness index shows that the NOMA scheme (with $k_{max}=2$ and $k_{max}=4$) provides fairer resource allocation for mMTC and URLLC devices compared to OMA. This is attributed to NOMA's ability to share resources, preventing some users from being completely deprived as can happen in OMA.
  • In terms of the number of users with satisfied rate requirements, the proposed NOMA scheme outperforms both OMA and the fast OFDM approaches. For instance, it can support more than 48 users, whereas the OMA scheme is limited by the number of sub-carriers.

Limitations and Future Work

  • NP-hard problem: The general optimization problem is NP-hard, requiring heuristic solutions. While the proposed heuristic is efficient, it may not always find the absolute optimal solution.
  • Single-cell scenario: The current model assumes a single-cell scenario, neglecting inter-cell interference. Future work could extend the model to multi-cell environments.
  • Fixed sub-carrier bandwidth: The paper assumes a fixed sub-carrier bandwidth. Investigating dynamic bandwidth allocation could be a future direction.
  • Further exploration of clustering strategies: While the proposed clustering is effective, exploring other clustering algorithms or dynamic clustering based on real-time network conditions could yield further improvements.

Summary updated

BibTeX
@article{shahini2019noma,
  author = {Ali Shahini and Nirwan Ansari},
  title = {{NOMA Aided Narrowband IoT for Machine Type Communications With User Clustering}},
  journal = {IEEE Internet of Things Journal},
  year = {2019},
  doi = {10.1109/jiot.2019.2914947},
}
Figure