Analytical Modeling of Interference Aware Power Control for the Uplink of Heterogeneous Cellular Networks

TL;DR

Stochastic geometry is used to analyze a power control mechanism that keeps the interference generated by each MT under a given threshold and reveals that such an interference aware method can reduce the average power consumption and increase the average spectral efficiency at once.

Abstract

Inter-cell interference is one of the main limiting factors in current Heterogeneous Cellular Networks (HCNs). Uplink Fractional Power Control (FPC) is a well known method that aims to cope with such limiting factor as well as to save battery live. In order to do that, the path losses associated with Mobile Terminal (MT) transmissions are partially compensated so that a lower interference is leaked towards neighboring cells. Classical FPC techniques only consider a set of parameters that depends on the own MT transmission, like desired received power at the Base Station (BS) or the path loss between the MT and its serving BS, among others. Contrary to classical FPC, in this paper we use stochastic geometry to analyze a power control mechanism that keeps the interference generated by each MT under a given threshold. We also consider a maximum transmitted power and a partial compensation of the path loss. Interestingly, our analysis reveals that such Interference Aware (IA) method can reduce the average power consumption and increase the average spectral efficiency. Additionally, the variance of the interference is reduced, thus improving the performance of Adaptive Modulation and Coding (AMC) since the interference can be better estimated at the MT.

Figures (7)

• Figure 1: Homogeneous network realization in $[-5000, 5000]^2$ m$^2$ showing with green color positions related to non truncated MTs. Yellow and blue colors are associated with MTs truncated by $i_0$ and $p_\mathrm{max}$ respectively. The simulation parameters are $\lambda^{(1)}=2$ BS/km$^2$, $i_0=-90$ dBm, $p_\mathrm{max} = 30$ dBm, $p_0=-70$ dBm and $\epsilon=1$.
• Figure 2: Average transmitted power versus $i_0$ for IAFPC and non IA with $p_\mathrm{max} \rightarrow \infty$ and $p_\mathrm{max} = 5$ dBm.
• Figure 3: (a) Mean of the interference versus $i_0$ for IAFPC and non IA with $p_\mathrm{max} \rightarrow \infty$ and $p_\mathrm{max} = 5$ dBm. (b) Variance of the interference versus $i_0$ for IAFPC and non IA with $p_\mathrm{max} \rightarrow \infty$ and $p_\mathrm{max} = 5$ dBm.
• Figure 4: Laplace transform of the analytic interference term $I$ given by (\ref{['eq:I']}) conditioned on $\mathcal{X}^{(1)}_{\mathrm{MT}_0}$ for IAFPC and its approximation by $\hat{I}$.
• Figure 5: (a) ccdf of the SINR for the typical MT using IAFPC method with $\epsilon=1$ and $i_0=\{-120, -60\}$ dBm. (b) ccdf of the SINR in the low $i_0$ regime given by (\ref{['eq:Low i0 min path ccdf SINR']}) using minimum path loss association.

Theorems & Definitions (14)

• Lemma 1
• proof
• Lemma 2
• proof
• Lemma 3
• proof
• Proposition 1
• proof
• Proposition 2
• proof