Models and Information Rates for Wiener Phase Noise Channels

TL;DR

This paper analyzes the capacity of Wiener phase noise channels using a multi-sample receiver, showing that oversampling improves information rates and capacity pre-log at high SNR, with amplitude modulation providing significant gains.

Contribution

It introduces a discrete-time model for Wiener phase noise channels with oversampling, demonstrating improved capacity pre-logs and analyzing the benefits of amplitude and phase modulation.

Findings

Oversampling at high SNR increases achievable rates.

Amplitude modulation achieves a pre-log of at least 1/2.

Numerical simulations confirm the benefits of oversampling for phase noise.

Abstract

A waveform channel is considered where the transmitted signal is corrupted by Wiener phase noise and additive white Gaussian noise. A discrete-time channel model that takes into account the effect of filtering on the phase noise is developed. The model is based on a multi-sample receiver, i.e., an integrate-and-dump filter whose output is sampled at a rate higher than the signaling rate. It is shown that, at high Signal-to-Noise Ratio (SNR), the multi-sample receiver achieves a rate that grows logarithmically with the SNR if the number of samples per symbol (oversampling factor) grows with the cubic root of the SNR. Moreover, the pre-log factor is at least 1/2 and can be achieved by amplitude modulation. For an approximate discrete-time model of the multi-sample receiver, the capacity pre-log at high SNR is shown to be at least 3/4 if the number of samples per symbol grows with the square root of the SNR. The analysis shows that phase modulation achieves a pre-log of at least 1/4 while amplitude modulation still achieves a pre-log of 1/2. This is strictly greater than the capacity pre-log of the (approximate) discrete-time Wiener phase noise channel with only one sample per symbol, which is 1/2. Numerical simulations are used to compute lower bounds on the information rates achieved by the multi-sample receiver. The simulations show that oversampling is beneficial for both strong and weak phase noise at high SNR. In fact, the information rates are sometimes substantially larger than when using commonly-used approximate discrete-time models.