Phase noise suppression in multi-gigabit millimeter wave systems

Abstract

Due to the potential of supporting multi-gigabit data rate, millimeter-wave (MMW) communication is deemed as a key technology for next-generation WLAN and 5G cellular systems. To fully realize the potential of MMW communication, we have to overcome some difficulties, such as phase noise (PHN). The immature CMOS fabrication incurs a relatively large PHN to MMW systems. Without suppression, the PHN can severely degrade the system performance. In this paper, we present a novel PHN suppression scheme for MMW single carrier frequency domain equalization (SC-FDE) systems. The key to suppress PHN is to estimate it as accurately as possible. Capitalizing on the statistical sparsity of MMW PHN in the basis formed by the eigenvectors of PHN autocorrelation matrix, we formulate the PHN estimation into the framework of compressive sensing (CS). In particular, we modify the sparse Bayesian leaning (SBL) strategy to estimate the involved PHN using the data decision available. The obtained PHN estimate is used to compensate for the PHN effect, followed by FDE operation and a re-demodulation of data symbols. This process iterates several times such that we can significantly improve the BER performance. Thanks to the modified SBL adopted, the proposed scheme can not only effectively mitigate the adverse effect of PHN, thereby leading to better performance, but also bear a low computation complexity.