Estimation of Oscillator Phase Noise for MPSK-based Commu- nication Systems over AWGN Channels

This work describes several methods to handle oscillator phase noise in single carrier systems operating over AWGN channels. The modulation type used is M -ary PSK (MPSK) with/without differential encoding. The oscillator phase noise is modeled as a Wiener-Lévy process. Data Aided (DA), Non-Data Aided (NDA) and Decision Directed (DD) maximum likelihood estimators are derived and a Pilot Symbol Assisted Modulation (PSAM) scheme using Wiener interpolation (WI) is explained. The PSAM scheme using Wiener interpolation, which can be applied for PSK modulation, is suitable for low signal-to-noise ratios (SNR ≤ 6dB). However, applying this scheme will lower the data throughput of the communication system. The NDA and DD techniques, developed from the DA estimator, can either observe received arguments or received complex phasors. The argument method, which relies on a phase unwrapping operation, is suitable for high SNRs only. The complex phasor method is suitable for medium SNRs and slowly time-varying phase noise processes. The NDA data dependency removal technique has a simple implementation. However, additional noise terms and ambiguous phase estimates are severe drawbacks associated with this method. DD estimators use more memory than NDA estimators and decision errors degrade the estimation process. Simulations show that the DD estimator has slightly better performance than the NDA estimator. A robust communication system, using NDA or DD oscillator phase noise compensation, thus requires differential encoded M -ary PSK (DMPSK) modulation to work properly. The design proposal for oscillator phase noise compensation schemes is summarized in the table below. Low SNR ≤ 6 dB Medium SNR High SNR PSAM WI NDA or DD – Complex phasor method NDA or DD – Argument method MPSK modulation DMPSK modulation DMPSK modulation Implementation issues should first be considered when deciding on which method to use.