[Paper Review] Exploiting NOMA for UAV Communications in Large-Scale Cellular Networks

Tianwei Hou, Yuanwei Liu+3 more|arXiv (Cornell University)|Feb 5, 2019

UAV Applications and Optimization41 references4 citations

TL;DR

This paper proposes a NOMA-assisted UAV communication framework in large-scale cellular networks, leveraging stochastic geometry to model UAV and user distributions. It introduces UAV-centric and user-centric strategies for offloading and emergency communications, deriving exact coverage probability expressions under imperfect SIC, showing that the ipSIC coefficient is a dominant factor in performance, with NOMA significantly outperforming OMA when power allocation and rates are optimized.

ABSTRACT

This paper advocates a pair of strategies in non-orthogonal multiple access (NOMA) in unmanned aerial vehicles (UAVs) communications, where multiple UAVs play as new aerial communications platforms for serving terrestrial NOMA users. A new multiple UAVs framework with invoking stochastic geometry technique is proposed, in which a pair of practical strategies are considered: 1) the UAV-centric strategy for offloading actions and 2) the user-centric strategy for providing emergency communications. In order to provide practical insights for the proposed NOMA assisted UAV framework, an imperfect successive interference cancelation (ipSIC) scenario is taken into account. For both UAV-centric strategy and user-centric strategy, we derive new exact expressions for the coverage probability. We also derive new analytical results for orthogonal multiple access (OMA) for providing a benchmark scheme. The derived analytical results in both user-centric strategy and UAV-centric strategy explicitly indicate that the ipSIC coefficient is a dominant component in terms of coverage probability. Numerical results are provided to confirm that i) for both user-centric strategy and UAV-centric strategy, NOMA assisted UAV cellular networks is capable of outperforming OMA by setting power allocation factors and targeted rate properly; and ii) the coverage probability of NOMA assisted UAV cellular framework is affected to a large extent by ipSIC coefficient, target rates and power allocations factors of paired NOMA users.

Motivation & Objective

• To address the need for efficient spectrum and energy use in UAV-based cellular networks.
• To enhance connectivity in disaster scenarios and temporary events through UAVs as aerial base stations.
• To improve spectral and energy efficiency by integrating NOMA into UAV networks.
• To model and analyze coverage performance in large-scale UAV networks using stochastic geometry.
• To compare NOMA and OMA performance under realistic imperfect SIC conditions.

Proposed method

• Proposes a large-scale UAV network model using a Poisson point process for UAV and user distribution.
• Applies stochastic geometry to derive Laplace transforms of interference from interfering UAVs and users.
• Models Nakagami-m fading for small-scale fading with LoS and non-LoS components.
• Uses Poisson Hole Process (PHP) to isolate the strongest interfering UAV at distance R.
• Derives exact coverage probability expressions for both UAV-centric and user-centric strategies under imperfect successive interference cancellation (ipSIC).
• Incorporates power-domain NOMA with superposition coding and SIC at the receiver, accounting for imperfect cancellation.

Experimental results

Research questions

• RQ1 How does NOMA improve coverage probability in UAV-assisted cellular networks compared to OMA?
• RQ2 What is the impact of the imperfect SIC coefficient on coverage performance in NOMA-based UAV networks?
• RQ3 How do power allocation factors and target rates affect the coverage probability in NOMA-assisted UAV systems?
• RQ4 What are the analytical performance limits of UAV-centric and user-centric strategies under stochastic geometry?
• RQ5 How does the proximity of the strongest interferer (at distance R) affect system coverage?

Key findings

• NOMA outperforms OMA in coverage probability when power allocation factors and target rates are properly optimized.
• The imperfect SIC coefficient is the dominant factor affecting coverage probability in both UAV-centric and user-centric strategies.
• Coverage probability is significantly degraded by imperfect SIC, highlighting the need for robust receiver design.
• The derived exact coverage probability expressions are valid for both NOMA and OMA, enabling direct performance comparison.
• The strongest interferer at distance R has a substantial impact, and its exclusion via Poisson Hole Process improves analytical tractability and accuracy.
• Numerical results confirm that NOMA with optimized parameters achieves higher spectral efficiency and better reliability than OMA in UAV networks.