Near-optimal quantization and linear network coding for relay networks

TL;DR

This paper introduces a quantization-based digital interface for Gaussian relay networks, demonstrating near-optimal capacity approximation and designing a robust linear coding strategy that is simple, channel-gain independent, and within a bounded gap of the network's capacity.

Contribution

It proposes a discrete network model obtained by quantization, showing its near-optimality for Gaussian networks, and develops a linear coding strategy that is robust and effective without channel gain knowledge.

Findings

Discrete network is within O(M^2) bits of Gaussian capacity.

Linear coding strategy achieves near-capacity rates.

Relays operate with simple quantization and linear transformations.

Abstract

We introduce a discrete network corresponding to any Gaussian wireless network that is obtained by simply quantizing the received signals and restricting the transmitted signals to a finite precision. Since signals in the discrete network are obtained from those of a Gaussian network, the Gaussian network can be operated on the quantization-based digital interface defined by the discrete network. We prove that this digital interface is near-optimal for Gaussian relay networks and the capacities of the Gaussian and the discrete networks are within a bounded gap of O(M^2) bits, where M is the number of nodes. We prove that any near-optimal coding strategy for the discrete network can be naturally transformed into a near-optimal coding strategy for the Gaussian network merely by quantization. We exploit this by designing a linear coding strategy for the case of layered discrete relay networks. The linear coding strategy is near-optimal for Gaussian and discrete networks and achieves rates within O(M^2) bits of the capacity, independent of channel gains or SNR. The linear code is robust and the relays need not know the channel gains. The transmit and receive signals at all relays are simply quantized to binary tuples of the same length $n$ . The linear network code requires all the relay nodes to collect the received binary tuples into a long binary vector and apply a linear transformation on the long vector. The resulting binary vector is split into smaller binary tuples for transmission by the relays. The quantization requirements of the linear network code are completely defined by the parameter $n$, which also determines the resolution of the analog-to-digital and digital-to-analog convertors for operating the network within a bounded gap of the network's capacity. The linear network code explicitly connects network coding for wireline networks with codes for Gaussian networks.