Robust Phase Noise Compensation in OFDM via a Dual-Model Approach

Abstract

A low-complexity, decision-directed, iterative scheme is proposed for severe phase noise mitigation in OFDM modulation over frequency-selective as well as flat channels. The proposed approach relies on two modeling steps: first, an alternative model formulation in the time domain; second, a maximum likelihood (ML)-inspired estimator which incorporates SNR-maximizing concepts. The resulting scheme is shown to outperform already proposed approaches in terms of achievable error rate and to do so with comparatively reduced complexity.