User-Coupling Angle-Domain Adaptive Filtering Based Frequency Synchronization for Massive MIMO Multiuser Uplink

Yinghao Ge,Weile Zhang

doi:10.1109/access.2019.2929843

Yinghao Ge, Weile Zhang

Open Access

https://doi.org/10.1109/access.2019.2929843

Copy DOI

Abstract

In this paper, we develop a user-coupling angle-domain adaptive filtering (UC-ADAF)-based frequency synchronization method for the uplink of a massive multiple-input multiple-output (MIMO) multiuser network, in the presence of highly overlapping angle-of-arrival (AoA) regions among coexisting users. In order to deal with the extremely severe multiuser interference (MUI) among users with highly overlapping AoA region, we propose to couple them together and perform joint carrier frequency offset (CFO) estimation and data detection. For the coupled users, the UC angle-constraining matrix (UC-ACM) can inherently suppress the interference from non-overlapping users and the UC-ADAF vectors are adaptively optimized to further eliminate the MUI from slightly overlapping users. Moreover, the challenging problem of addressing the severe MUI caused by highly overlapping users is circumvented through joint estimation and detection. The effectiveness of the proposed UC-ADAF approach and its robustness against overlapping degree among coupled users and the total number of mutually overlapping users, as well as its superiority over the existing methods, are demonstrated by numerical results.

Highlights

Over the past few years, large-scale multiple-input multiple-output (MIMO) or massive MIMO systems have drawn exploding interests from both academia and industry [1]–[11]
The drawbacks of the adaptive filtering (ADAF) approach motivates us to design in this paper a user-coupling ADAF (UC-ADAF) based frequency synchronization approach to cope with the VOLUME 7, 2019 multi-carrier frequency offset (CFO) estimation of users with highly overlapping AoA region
Unlike the user-coupling angle-domain adaptive filtering (UC-ADAF), the user grouping (UG) groups all the 9 users to carry out the joint CFO estimation and data detection, since their AoA regions are contiguously overlapping without sufficient angular separation

Summary

INTRODUCTION

Over the past few years, large-scale multiple-input multiple-output (MIMO) or massive MIMO systems have drawn exploding interests from both academia and industry [1]–[11]. As the intragroup users exceed a certain number, the UG technique in [28] exhibits an underdetermination problem, thereby suffering from severe performance exacerbation Another method targeting at addressing the multi-CFO estimation issue in the presence of overlapping AoA regions among users is the angle-domain adaptive filtering (ADAF) approach in [29], [30], which can achieve two-stage MUI suppression. The drawbacks of the ADAF approach motivates us to design in this paper a user-coupling ADAF (UC-ADAF) based frequency synchronization approach to cope with the VOLUME 7, 2019 multi-CFO estimation of users with highly overlapping AoA region. Diag(x) is a diagonal matrix with x as the main diagonal; ⊗ denotes the Kronecker product; vec(·) is the vectorization operator; Cm×n defines the vector space of all m × n complex matrices; IQ is the Q × Q identity matrix

SYSTEM MODEL

UC-ACM AND UC-ADAF VECTORS

SIMULATION RESULTS

CONCLUSION