[Paper Review] Coverage and capacity scaling laws in downlink ultra-dense cellular networks

Van Minh Nguyen, Marios Kountouris|arXiv (Cornell University)|Feb 10, 2016

Advanced MIMO Systems Optimization17 references4 citations

TL;DR

This paper analyzes downlink coverage and capacity scaling laws in ultra-dense cellular networks using a Poisson point process model with multi-slope path loss and general fading distributions. It reveals that coverage and capacity exhibit an 'inverse U' shape with increasing density when near-field path loss is weak (β₀ < d) or fading is light-tailed, leading to performance degradation at high densities—highlighting fundamental limits of network densification beyond which further deployment harms performance.

ABSTRACT

Driven by new types of wireless devices and the proliferation of bandwidth-intensive applications, data traffic and the corresponding network load are increasing dramatically. Network densification has been recognized as a promising and efficient way to provide higher network capacity and enhanced coverage. Most prior work on performance analysis of ultra-dense networks (UDNs) has focused on random spatial deployment with idealized singular path loss models and Rayleigh fading. In this paper, we consider a more precise and general model, which incorporates multi-slope path loss and general fading distributions. We derive the tail behavior and scaling laws for the coverage probability and the capacity considering strongest base station association in a Poisson field network. Our analytical results identify the regimes in which the signal-to-interference-plus-noise ratio (SINR) either asymptotically grows, saturates, or decreases with increasing network density. We establish general results on when UDNs lead to worse or even zero SINR coverage and capacity, and we provide crisp insights on the fundamental limits of wireless network densification.

Motivation & Objective

• To investigate the fundamental performance limits of ultra-dense networks (UDNs) under realistic propagation conditions beyond idealized models.
• To address the gap in prior work by incorporating multi-slope path loss and general fading distributions instead of singular path loss and Rayleigh fading.
• To determine under what conditions network densification improves, saturates, or degrades SINR coverage and system capacity.
• To identify the regimes where coverage and capacity are maximized at finite network densities rather than monotonically improving with densification.

Proposed method

• Models base station locations as a homogeneous Poisson point process (PPP) in a d-dimensional plane.
• Introduces a multi-slope path loss model with distinct path loss exponents in near-field (β₀) and far-field (βₖ) regions.
• Uses general fading distributions, including regularly varying tails (e.g., α ∈ (0,1)), to model fast fading, shadowing, or composite fading.
• Analyzes the tail behavior of received signal power and SINR using asymptotic analysis and regular variation theory.
• Derives scaling laws for coverage probability and system capacity under strongest base station association.
• Applies Karamata’s theorem and monotone density theorem to characterize the asymptotic behavior of fading and interference distributions.

Experimental results

Research questions

• RQ1 How does multi-slope path loss affect the scaling behavior of coverage probability and capacity in ultra-dense networks?
• RQ2 Under what conditions does increasing network density lead to a decrease in SINR coverage and system capacity?
• RQ3 What is the impact of general fading distributions—especially those with regularly varying tails—on the performance scaling of UDNs?
• RQ4 When does network densification lead to performance saturation or degradation rather than improvement?
• RQ5 What are the fundamental limits of network densification in terms of achievable coverage and capacity?

Key findings

• When the near-field path loss exponent β₀ < d (network dimension), coverage and capacity exhibit an 'inverse U' shape with increasing network density, peaking at a finite density before degrading.
• If β₀ > d or fading is heavy-tailed (i.e., F̄ₘ ∈ R₋ᵅ with α ∈ (0,1)), coverage and capacity saturate at a finite upper bound as density increases.
• Regularly varying fading distributions (F̄ₘ ∈ R₋ᵅ) have the same detrimental effect on performance scaling as path loss singularities, leading to performance collapse at high densities.
• Bounded path loss (β₀ = 0) is a special case of β₀ < d, and in such cases, both coverage and capacity are maximized at a finite network density.
• All standard fading models (Rayleigh, lognormal, Gamma, and their composites) fall into the same class of fading distributions that cause coverage and capacity to peak at finite densities.
• When β₀ > 0 and fading is light-tailed (α₀ < α), coverage and capacity are limited by a constant bound that does not vanish even as density approaches infinity.