168 0

Downlink Pilot Reduction for Massive MIMO Systems via Compressed Sensing

Title
Downlink Pilot Reduction for Massive MIMO Systems via Compressed Sensing
Author
최준원
Keywords
Channel estimation; compressed sensing; downlink pilot allocation; massive multiple-input multiple-output (MIMO); orthogonal frequency division multiplexing (OFDM)
Issue Date
2015-11
Publisher
IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC
Citation
IEEE COMMUNICATIONS LETTERS, v. 19, NO 11, Page. 1889-1892
Abstract
This letter addresses a problem of downlink pilot allocation for massive multiple-input multiple-output (MIMO) systems. When a massive MIMO is employed in frequency division duplex (FDD) systems, significant amount of radio resources are dedicated to the transmission of downlink pilots. Such huge pilot overhead leads to a substantial loss in the maximum data throughput, which motivates us to reduce the number of pilots. In this letter, we propose a pilot reduction strategy based on compressed sensing techniques for orthogonal frequency division multiplexing systems. The pilots are randomly located in a low density manner over the time and frequency domain. To estimate the channels with such low density pilots, we propose a novel sparse channel estimation technique that exploits the common support of the consecutive channel impulse responses over the certain time duration. The evaluation shows that for a massive MIMO with 128 antennas, the proposed scheme achieves significant reduction of pilot overhead, while maintaining good channel estimation performance.
URI
http://ieeexplore.ieee.org/document/7229286/http://hdl.handle.net/20.500.11754/28871
ISSN
1089-7798; 1558-2558
DOI
10.1109/LCOMM.2015.2474398
Appears in Collections:
COLLEGE OF ENGINEERING[S](공과대학) > ELECTRICAL AND BIOMEDICAL ENGINEERING(전기·생체공학부) > Articles
Files in This Item:
There are no files associated with this item.
Export
RIS (EndNote)
XLS (Excel)
XML


qrcode

Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.

BROWSE