1Asymptotic analysis of precoded small cell networks

Sreenath Ramanath, Mérouane Debbah, Eitan Altman, Vinod Kumar2015

TL;DR

This paper investigates how interference between neighboring small cells impacts the data capacity of MIMO wireless networks. Using advanced random matrix theory, the authors derive formulas for the maximum achievable data rate, showing that interference significantly limits capacity gains, especially as networks grow. They also identify optimal configurations for network parameters to maximize capacity.

What Problem Does This Paper Address?

The proliferation of small cells (like picocells) is a key strategy for increasing wireless network capacity and improving user experience, especially for mobile broadband. Unlike traditional large "macro" cells, small cells are densely packed, leading to substantial "inter-cell interference" – signals from one cell disrupting communications in another. This interference can negate the benefits of using multiple antennas at base stations (MIMO), which are typically used to boost data rates. The paper aims to quantify this impact of inter-cell interference on the achievable sum-rate capacity in MIMO small cell networks, particularly when base stations use linear precoding techniques.

Methods

The core of the paper's methodology lies in the application of random matrix theory (RMT). RMT is a powerful mathematical framework for analyzing the behavior of large systems with many random components, which is highly relevant to complex wireless networks. The authors leverage RMT tools, specifically the Stieltjes transform, to derive analytical expressions for the ergodic sum-rate capacity. This allows them to study the system's behavior in the asymptotic regime, where the number of antennas at the base station (M) and the number of users (K) become very large, while their ratio (β = M/K) is kept constant.

They focus on channel inversion (CI) precoding, a linear technique that aims to eliminate inter-user interference within a cell. The analysis extends from a single-cell scenario to a multi-cell scenario, incorporating interference from adjacent cells. A key aspect of their work is identifying an optimal ratio β* (M/K) that maximizes the sum-rate capacity for a given signal-to-noise ratio (SNR) and interference level.

Datasets

Not discussed in the paper. The paper relies on theoretical derivations and simulations, not on specific real-world datasets.

Key Insights and Ideas

  • Interference is a major bottleneck: The paper strongly demonstrates that inter-cell interference significantly diminishes the capacity gains achievable by MIMO in small cell networks. This effect is more pronounced than in scenarios with less interference.
  • Sub-linear capacity growth: As interference increases, the sum-rate capacity tends to grow sub-linearly with the number of antennas, meaning you get diminishing returns. Simply adding more antennas doesn't proportionally increase capacity when interference is high.
  • Optimal network configuration exists: There's an optimal balance between the number of antennas at the base station (M) and the number of users (K) – represented by the ratio β = M/K – that maximizes the sum-rate capacity. This optimal ratio (β*) is dependent on the interference level and transmit power.
  • Interference management is crucial: The paper highlights that while interference is detrimental, the proposed analysis and the effectiveness of linear precoding suggest that interference management techniques can still be highly effective even in dense small cell environments.
  • RMT provides accurate predictions: The asymptotic results derived using RMT are shown to be good approximations of the actual system performance, even for finite numbers of antennas and users, validating the theoretical approach.

Key Results

The paper derives analytical expressions for the ergodic sum-rate capacity in both single-cell and multi-cell MIMO networks with channel inversion precoding.

  • Single-cell: The sum-rate capacity is shown to be $R_{ci} = K \log(1 + \frac{\rho\beta}{\beta S_{H/H^H}(0)})$, where $\rho$ is related to transmit power and $S_{H/H^H}(0)$ is a term derived from RMT. An optimal $\beta^*$ is found by solving an implicit equation.
  • Multi-cell: With interference factor $\gamma$, the sum-rate capacity is given by $R_{ci} = K \log(1 + \frac{\rho\beta}{\beta S_{H/H^H}(0) + 2\gamma\rho})$. The paper also derives an optimal $\beta^*$ for the multi-cell case, which decreases with increasing interference.
  • Simulations: Figures 2-9 visually confirm these findings. For instance, Figure 4 shows that the sum-rate at the optimal $\beta^*$ saturates more quickly with higher interference. Figure 7 illustrates the sensitivity of the rate to interference, showing a steep drop initially. Figures 8 and 9 demonstrate that capacity growth becomes sub-linear with increasing antennas (M) when interference is present.

Limitations and Future Work

  • Assumptions: The analysis makes several simplifying assumptions, including:
  • Uncorrelated base station antennas.
  • Wyner-type cellular models (simplified interference structure).
  • Users experiencing similar average received signal power (no deep shadow fades).
  • Perfect channel state information (CSIT) at the transmitter.
  • Rayleigh fading and path-loss.
  • Future Work: The authors suggest that further investigation into interference coordination schemes in small cell networks would be valuable. They also imply that exploring more complex channel models and interference scenarios could lead to more refined capacity analyses. The paper's findings lay the groundwork for understanding how to design and optimize future dense wireless networks.

Summary updated

Abstract

In this paper, we study precoded MIMO based small cell networks. We derive the theoretical sum-rate capacity, when multi-antenna base stations transmit precoded information to its multiple single-antenna users in the presence of inter-cell interference from neighboring cells. Due to an interference limited scenario, increasing the number of antennas at the base stations does not yield necessarily a linear increase of the capacity. We assess exactly the effect of multi-cell interference on the capacity gain for a given interference level. We use recent tools from random matrix theory to obtain the ergodic sum-rate capacity, as the number of antennas at the base station, number of users grow large. Simulations confirm the theoretical claims and also indicate that in most scenarios the asymptotic derivations applied to a finite number of users give good approximations of the actual ergodic sum-rate capacity. Index Terms—Cellular networks; MIMO; Small cells; random matrix theory; linear precoding. I.
BibTeX
@misc{ramanath2015asymptotic,
  author = {Sreenath Ramanath and Mérouane Debbah and Eitan Altman and Vinod Kumar},
  title = {{1Asymptotic analysis of precoded small cell networks}},
  year = {2015},
}
Figure