Analysis and Management of Interference in 5G Reverse TDD HetNets

Luca Sanguinetti, Aris L. Moustakas, Mérouane Debbah5GU 2014 · DOI

Abstract

This work analyzes a heterogeneous network (HetNet) comprising a macro base station (BS) equipped with a large number of antennas and an overlaid dense tier of small cell access points (SCAs) utilizing wireless backhaul. By employing a "reverse TDD" protocol and large system analysis, the authors derive closed-form expressions for transmit power and identify a critical threshold where imperfect channel knowledge renders the system physically unable to meet user rate demands.

Introduction

As 5G networks densify to meet increasing data demands, managing interference between macro-cells and small-cells becomes paramount. Traditional wired backhaul for every small cell is often cost-prohibitive and geographically impractical. This study addresses two primary challenges in such environments: the coordination of inter-tier interference and the impact of imperfect Channel State Information (CSI) resulting from user mobility.

System Model and the Reverse TDD Protocol

The proposed architecture features a Massive MIMO Macro BS providing wireless backhaul to a dense layer of SCAs. To mitigate the complex interference patterns inherent in this two-tier setup, the researchers introduce a Reverse Time Division Duplexing (TDD) protocol.

In this framework, the transmission directions are interleaved: when the Macro BS is in Downlink (DL) mode (transmitting to Macro UEs and SCAs), the SCAs are in Uplink (UL) mode (receiving from Small Cell UEs). Conversely, when the Macro BS is in UL mode, the SCAs operate in DL mode. This reversal allows the Macro BS to leverage its massive antenna array to beamform signals precisely, effectively nullifying interference between the tiers.

Reverse TDD protocol illustrating the interleaving of UL/DL phases to manage inter-tier interference.

Large System Analysis and Power Divergence

The paper utilizes Large System Analysis and Random Matrix Theory to evaluate performance in the asymptotic regime, where the number of antennas ($N$) and users ($K$) grow proportionally large. This approach yields concise mathematical formulas for power consumption, which are otherwise intractable for specific, finite deployments.

A significant theoretical contribution is the identification of the Power Divergence Threshold. The analysis reveals that the required transmit power is highly sensitive to channel uncertainty ($\tau^2$), which arises from user mobility and estimation errors. As shown in the performance plots, there is a "critical value" for $\tau^2$; if the CSI becomes too stale or the target rate too high, the power required to maintain the link diverges to infinity.

Average transmit power for MUEs and SCAs versus target rate, demonstrating the power divergence threshold as CSI error ($\tau^2$) increases.

Key Findings and Performance Metrics

  • Wireless Backhaul Viability: The study demonstrates that using wireless backhaul for SCAs does not significantly degrade the power efficiency of macro users compared to a traditional wired backhaul architecture.
  • Spectral and Power Efficiency: Numerical results indicate that for a target rate of 2 bit/s/Hz, the BS requires only 0.1 Watts of transmit power to serve the entire network, achieving an aggregate area throughput of approximately 5.12 Gbit/s/km².
  • Mobility Constraints: Simulations show that at a mobility of 50 km/h ($\tau^2 = 0.3$), the system becomes infeasible if target rates exceed 1.5 bit/s/Hz, highlighting a fundamental physical limit of the architecture.

Conclusion

The integration of Massive MIMO with a Reverse TDD protocol provides a robust solution for managing interference in dense 5G HetNets. While wireless backhaul is shown to be efficient, the existence of a power divergence threshold underscores the critical importance of CSI accuracy. For scenarios with high user mobility, the results suggest that orthogonal resource allocation may be necessary to maintain system stability.

Summary updated

BibTeX
@inproceedings{sanguinetti2014analysis,
  author = {Luca Sanguinetti and Aris L. Moustakas and Mérouane Debbah},
  title = {{Analysis and Management of Interference in 5G Reverse TDD HetNets}},
  booktitle = {5GU},
  year = {2014},
  doi = {10.4108/icst.5gu.2014.258124},
}
Figure