Interference Statistics and Capacity Analysis for Uplink Transmission in Two-Tier Small Cell Networks: A Geometric Probability Approach

Hina Tabassum, Zaher Dawy, Ekram Hossain, Mohamed-Slim Alouini

TL;DR

This paper introduces a novel geometric probability framework to analyze interference and capacity in two-tier cellular networks (macrocells with overlaid femtocells). It provides a more tractable analytical model for interference statistics, considering various real-world factors, and offers insights into optimizing spectrum access strategies for improved network performance.

What Problem Does This Paper Address?

Modern cellular networks are increasingly deploying "small cells" like femtocells to boost capacity and coverage. However, this dense deployment, often unplanned, leads to significant interference, which degrades overall network performance. Existing analytical models often struggle to capture the complex interplay of factors like random femtocell placement, different spectrum access methods (shared vs. dedicated), varying traffic loads, and realistic channel conditions (shadowing and fading). This paper aims to bridge this gap by developing a more accurate and tractable analytical framework to understand and mitigate interference in these two-tier networks.

Methods

The authors propose a novel framework that leverages geometric probability theory to model the spatial distribution of interferers and their distances to receivers. Key techniques include:

  • Random Line Picking Theory: Used to model distances for uplink interference from users within the same macrocell.
  • Distance Distribution in Concentric Circles: Employed to approximate cross-tier interference from users in adjacent macrocells.
  • Approximation of Generalized-K Fading: The complex Generalized-K composite fading distribution is approximated by the more tractable Gamma distribution using moment matching, simplifying calculations while retaining accuracy.
  • Moment Generating Function (MGF) Approach: Instead of direct convolution of probability density functions (PDFs), the MGF of the cumulative interference is derived. This significantly reduces computational complexity, allowing for single-integral capacity calculations.
  • Hybrid Macrocell Model: The framework combines a traditional grid-based model for macrocells with the random deployment of femtocells, offering a more realistic representation than purely grid-based or purely Poisson Point Process (PPP) models.

Datasets

Not discussed in the paper. The authors rely on analytical derivations and validate their results using Monte-Carlo simulations.

Key Insights and Ideas

  • Geometric Probability is Key: Using geometric probability for distance distributions provides a powerful and flexible way to model interference in randomly deployed small cell networks.
  • Gamma Approximation is Tractable: Approximating Generalized-K fading with Gamma distribution significantly simplifies the analysis without substantial loss of accuracy.
  • MGF Simplifies Capacity: The MGF-based approach for cumulative interference is crucial for deriving tractable capacity expressions, avoiding complex convolutions.
  • Dedicated vs. Shared Spectrum Trade-offs: The paper highlights that shared spectrum access can offer higher femtocell capacity in high traffic scenarios, even at the cost of reduced macrocell capacity, while dedicated spectrum offers more predictable performance.
  • Co-tier Interference Matters: For dense femtocell deployments and heavy traffic, co-tier interference cannot be ignored and significantly degrades per-subcarrier capacity.
  • Grid vs. Random Deployment Models: Circular grid-based models for femtocell deployment can significantly overestimate or underestimate interference compared to random deployment models.

Key Results

  • The paper derives closed-form expressions for the statistics (PDF and MGF) of co-tier and cross-tier interference in two-tier networks.
  • Analytical expressions for per-subcarrier and aggregate ergodic capacity are derived for both dedicated and shared spectrum access scenarios.
  • Monte-Carlo simulations validate the accuracy of the derived analytical expressions across various scenarios, including different traffic loads, deployment densities, and spectral reuse strategies.
  • Numerical results demonstrate that while aggregate femtocell capacity increases with density and spectral reuse, per-subcarrier capacity degrades due to co-tier interference.
  • Shared spectrum access is shown to be a beneficial strategy in high traffic load and dense femtocell deployment scenarios, despite potential reductions in macrocell capacity.

Limitations and Future Work

  • Approximation of Generalized-K: While the Gamma approximation is tractable, it's still an approximation. The paper notes that exact expressions for Generalized-K fading could be derived at the cost of increased computational complexity.
  • Downlink Analysis: The paper primarily focuses on uplink transmission. While it mentions that the framework can be extended to downlink, it highlights the need for careful scaling of interference and signal variables due to potential differences in transmission powers.
  • Specific Network Topologies: The analysis assumes a grid-based macrocell structure and random femtocell placement. Future work could explore other macrocell topologies or more complex femtocell deployment patterns.
  • Interference from Neighboring Macrocells: The paper notes that interference from femtocells in distant macrocell tiers is often negligible, but a more comprehensive analysis could incorporate these effects for completeness.
  • Dynamic Traffic and Mobility: The current model assumes static traffic loads. Extending the framework to handle dynamic traffic variations and user mobility would be a valuable future direction.

Summary updated

Abstract

This paper presents a novel framework to derive the statistics of the interference considering dedicated and shared spectrum access for uplink transmission in two-tier small cell networks such as the macrocell-femtocell networks. The framework exploits the distance distributions from geometric probability theory to characterize the uplink interference while considering a traditional grid-model set-up for macrocells along with the randomly deployed femtocells. The derived expressions capture the impact of path-loss, composite shadowing and fading, uniform and non-uniform traffic loads, spatial distribution of femtocells, and partial and full spectral reuse among femtocells. Considering dedicated spectrum access, first, we derive the statistics of co-tier interference incurred at both femtocell and macrocell base stations (BSs) from a single interferer by approximating generalized-K composite fading distribution with the tractable Gamma distribution. We then derive the distribution of the number of interferers considering partial spectral reuse and moment generating function (MGF) of the cumulative interference for both partial and full spectral reuse scenarios. Next, we derive the statistics of the cross-tier interference at both femtocell and macrocell BSs considering shared spectrum access. Finally, we utilize the derived expressions to analyze the capacity in both dedicated and shared spectrum access scenarios. The derived expressions are validated by the Monte Carlo simulations. Numerical results are generated to assess the feasibility of shared and dedicated spectrum access in femtocells under varying traffic load and spectral reuse scenarios.
BibTeX
@article{tabassum2014interference,
  author = {Hina Tabassum and Zaher Dawy and Ekram Hossain and Mohamed-Slim Alouini},
  title = {{Interference Statistics and Capacity Analysis for Uplink Transmission in Two-Tier Small Cell Networks: A Geometric Probability Approach}},
  journal = {IEEE Trans. Wireless Communications},
  year = {2014},
  url = {https://openreview.net/forum?id=9Oy3GkySQSx},
}
Figure