[Paper Review] On Uplink-Downlink Duality for Cellular IA

Vasileios Ntranos, Mohammad Ali Maddah-Ali+1 more|arXiv (Cornell University)|Jul 14, 2014

Advanced MIMO Systems Optimization16 references4 citations

TL;DR

This paper establishes uplink-downlink duality for cellular interference alignment (IA) in a hexagonal cellular network with local backhaul links. It proposes a downlink scheme where base stations exchange quantized dirty-paper precoded signals to enable successive interference cancellation, proving that any degrees of freedom (DoF) achievable via one-shot linear IA in the uplink are also achievable in the downlink using a dual precoding scheme with reversed encoding order.

ABSTRACT

In our previous work we considered the uplink of a hexagonal cellular network topology and showed that linear "one-shot" interference alignment (IA) schemes are able to achieve the optimal degrees of freedom (DoFs) per user, under a decoded-message passing framework that allows base-stations to exchange their own decoded messages over local backhaul links. In this work, we provide the dual framework for the downlink of cellular networks with the same backhaul architecture, and show that for every "one-shot" IA scheme that can achieve $d$ DoFs per user in the uplink, there exists a dual "one-shot" IA scheme that can achieve the same DoFs in the downlink. To enable "Cellular IA" for the downlink, base-stations will now use the same local backhaul links to exchange quantized versions of the dirty-paper precoded signals instead of user messages.

Motivation & Objective

• To extend the one-shot interference alignment framework from the uplink to the downlink in cellular networks with local backhaul links.
• To address the fundamental question of whether similar degrees of freedom (DoF) can be achieved in the downlink as in the uplink under the same backhaul architecture.
• To develop a dual downlink transmission scheme that enables interference alignment using quantized precoded signals instead of decoded messages.
• To establish a formal duality between uplink and downlink interference alignment by reversing the encoding order and swapping beamforming matrices.
• To demonstrate that dirty-paper coding (DPC) enables effective interference pre-cancellation in the downlink, unlike prior linear precoding schemes.

Proposed method

• Proposes a successive encoding scheme in the downlink where base stations (BSs) quantize and share their dirty-paper precoded signals over local backhaul links.
• Uses a reversed encoding order $\overline{\pi}$ compared to the uplink decoding order $\pi$, enabling sequential interference cancellation.
• Applies dirty-paper coding (DPC) at each base station to pre-cancel known interference from previously encoded signals.
• Establishes duality by mapping uplink beamforming matrices $\mathbf{V}_v, \mathbf{U}_u$ to downlink transmit and receive beamformers $\overline{\mathbf{V}}_v = \mathbf{U}_v, \overline{\mathbf{U}}_u = \mathbf{V}_u$.
• Leverages channel reciprocity: downlink channel matrices $\overline{\mathbf{H}}_{vu} = \mathbf{H}_{uv}^H$ are the Hermitian transpose of uplink channels.
• Demonstrates that the interference alignment (IA) conditions in the uplink are preserved in the downlink under the dual framework, ensuring DoF equivalence.

Experimental results

Research questions

• RQ1 Can the same degrees of freedom (DoF) achieved by one-shot linear interference alignment in the uplink be achieved in the downlink under the same backhaul constraints?
• RQ2 What is the appropriate downlink transmission strategy that enables interference alignment when base stations can only exchange quantized signals over local backhaul links?
• RQ3 How does dirty-paper coding (DPC) enable effective interference pre-cancellation in the downlink under a successive encoding framework?
• RQ4 What is the precise duality relationship between uplink decoding order and downlink encoding order in interference alignment schemes?
• RQ5 Can the duality principle from uplink IA be extended to cellular networks with a hexagonal topology and local backhaul cooperation?

Key findings

• Any degrees of freedom $\{d_v, v \in \mathcal{V}\}$ achievable via one-shot linear IA in the uplink are also achievable in the downlink using the dual scheme.
• The downlink achieves the same DoF as the uplink by using the reverse encoding order $\overline{\pi}$ and swapping beamforming matrices: $\overline{\mathbf{V}}_v = \mathbf{U}_v$, $\overline{\mathbf{U}}_u = \mathbf{V}_u$.
• The interference alignment condition $\mathbf{U}_u^H \mathbf{H}_{uv} \mathbf{V}_v = 0$ in the uplink is transformed into $\overline{\mathbf{U}}_u^H \overline{\mathbf{H}}_{uv} \overline{\mathbf{V}}_v = 0$ in the downlink via channel reciprocity.
• The signal-to-interference-plus-noise ratio (SINR) in the downlink is preserved such that each user achieves $d_v$ degrees of freedom, as the interference terms are nullified by DPC.
• The scheme does not require message sharing or quantized received signal exchange; instead, it exchanges quantized dirty-paper precoded signals, enabling pre-cancellation of known interference.
• The duality holds under the same network topology $\mathcal{G}(\mathcal{V}, \mathcal{E})$ and backhaul rate constraints, proving equivalence in DoF performance between uplink and downlink under the proposed framework.