Pre-DFT processing using eigenanalysis for coded OFDM with multiple receive antennas

Defeng Huang, Khaled Letaief

Research output: Contribution to journalArticle

12 Citations (Scopus)

Abstract

In broadband wireless communications, coded orthogonal frequency-division multiplexing (OFDM) can be used with multiple receive antennas to achieve both frequency diversity and space diversity. In this scenario, the optimal approach is subcarrier-based space combining. However, such an approach is quite complex, because multiple discrete Fourier transform (DFT) blocks, each per receive antenna, are used. In this paper, we propose a pre-DFT processing scheme based upon eigenanalysis. In the proposed scheme, the received signals are weighted and combined both before and after the DFT processing. As a result, the required number of DFT blocks can be significantly reduced. With perfect weighting coefficients, the margin of the performance improvement decreases along with the increase of the number of DFT blocks, thus enabling effective performance and complexity tradeoff. To achieve a maximum average pairwise codeword distance, it will be shown that the maximum number of DFT blocks required is equal to the minimum of the number of receive antennas and the number of distinct paths in the channel. When the number of distinct paths is larger than the number of receive antennas and with a smaller number of DFT blocks, extensive simulation results will also show that near-optimal performance can still be achieved for most channels. Finally, in an OFDM system with differential modulation, we use a signal covariance matrix to obtain the weighting coefficients before the DFT processing. In this case, simulation results will demonstrate that the performance of the proposed scheme can be better than subcarrier-based space combining, but with much lower complexity.

Original languageEnglish
Pages (from-to)2019-2027
Number of pages9
JournalIEEE Transactions on Communications
Volume52
Issue number11
DOIs
Publication statusPublished - Nov 2004
Externally publishedYes

Fingerprint

Discrete Fourier transforms
Orthogonal frequency division multiplexing
Antennas
Processing
Covariance matrix
Modulation
Communication

Keywords

  • Broadband wireless communications
  • Diversity
  • Eigen-analysis
  • Multiple receive antennas
  • Orthogonal frequency-division multiplexing (OFDM)

ASJC Scopus subject areas

  • Electrical and Electronic Engineering
  • Computer Networks and Communications

Cite this

Pre-DFT processing using eigenanalysis for coded OFDM with multiple receive antennas. / Huang, Defeng; Letaief, Khaled.

In: IEEE Transactions on Communications, Vol. 52, No. 11, 11.2004, p. 2019-2027.

Research output: Contribution to journalArticle

@article{9f67b708dbae4b81ad0c1f07ba18d2eb,
title = "Pre-DFT processing using eigenanalysis for coded OFDM with multiple receive antennas",
abstract = "In broadband wireless communications, coded orthogonal frequency-division multiplexing (OFDM) can be used with multiple receive antennas to achieve both frequency diversity and space diversity. In this scenario, the optimal approach is subcarrier-based space combining. However, such an approach is quite complex, because multiple discrete Fourier transform (DFT) blocks, each per receive antenna, are used. In this paper, we propose a pre-DFT processing scheme based upon eigenanalysis. In the proposed scheme, the received signals are weighted and combined both before and after the DFT processing. As a result, the required number of DFT blocks can be significantly reduced. With perfect weighting coefficients, the margin of the performance improvement decreases along with the increase of the number of DFT blocks, thus enabling effective performance and complexity tradeoff. To achieve a maximum average pairwise codeword distance, it will be shown that the maximum number of DFT blocks required is equal to the minimum of the number of receive antennas and the number of distinct paths in the channel. When the number of distinct paths is larger than the number of receive antennas and with a smaller number of DFT blocks, extensive simulation results will also show that near-optimal performance can still be achieved for most channels. Finally, in an OFDM system with differential modulation, we use a signal covariance matrix to obtain the weighting coefficients before the DFT processing. In this case, simulation results will demonstrate that the performance of the proposed scheme can be better than subcarrier-based space combining, but with much lower complexity.",
keywords = "Broadband wireless communications, Diversity, Eigen-analysis, Multiple receive antennas, Orthogonal frequency-division multiplexing (OFDM)",
author = "Defeng Huang and Khaled Letaief",
year = "2004",
month = "11",
doi = "10.1109/TCOMM.2004.836594",
language = "English",
volume = "52",
pages = "2019--2027",
journal = "IEEE Transactions on Communications",
issn = "0096-1965",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
number = "11",

}

TY - JOUR

T1 - Pre-DFT processing using eigenanalysis for coded OFDM with multiple receive antennas

AU - Huang, Defeng

AU - Letaief, Khaled

PY - 2004/11

Y1 - 2004/11

N2 - In broadband wireless communications, coded orthogonal frequency-division multiplexing (OFDM) can be used with multiple receive antennas to achieve both frequency diversity and space diversity. In this scenario, the optimal approach is subcarrier-based space combining. However, such an approach is quite complex, because multiple discrete Fourier transform (DFT) blocks, each per receive antenna, are used. In this paper, we propose a pre-DFT processing scheme based upon eigenanalysis. In the proposed scheme, the received signals are weighted and combined both before and after the DFT processing. As a result, the required number of DFT blocks can be significantly reduced. With perfect weighting coefficients, the margin of the performance improvement decreases along with the increase of the number of DFT blocks, thus enabling effective performance and complexity tradeoff. To achieve a maximum average pairwise codeword distance, it will be shown that the maximum number of DFT blocks required is equal to the minimum of the number of receive antennas and the number of distinct paths in the channel. When the number of distinct paths is larger than the number of receive antennas and with a smaller number of DFT blocks, extensive simulation results will also show that near-optimal performance can still be achieved for most channels. Finally, in an OFDM system with differential modulation, we use a signal covariance matrix to obtain the weighting coefficients before the DFT processing. In this case, simulation results will demonstrate that the performance of the proposed scheme can be better than subcarrier-based space combining, but with much lower complexity.

AB - In broadband wireless communications, coded orthogonal frequency-division multiplexing (OFDM) can be used with multiple receive antennas to achieve both frequency diversity and space diversity. In this scenario, the optimal approach is subcarrier-based space combining. However, such an approach is quite complex, because multiple discrete Fourier transform (DFT) blocks, each per receive antenna, are used. In this paper, we propose a pre-DFT processing scheme based upon eigenanalysis. In the proposed scheme, the received signals are weighted and combined both before and after the DFT processing. As a result, the required number of DFT blocks can be significantly reduced. With perfect weighting coefficients, the margin of the performance improvement decreases along with the increase of the number of DFT blocks, thus enabling effective performance and complexity tradeoff. To achieve a maximum average pairwise codeword distance, it will be shown that the maximum number of DFT blocks required is equal to the minimum of the number of receive antennas and the number of distinct paths in the channel. When the number of distinct paths is larger than the number of receive antennas and with a smaller number of DFT blocks, extensive simulation results will also show that near-optimal performance can still be achieved for most channels. Finally, in an OFDM system with differential modulation, we use a signal covariance matrix to obtain the weighting coefficients before the DFT processing. In this case, simulation results will demonstrate that the performance of the proposed scheme can be better than subcarrier-based space combining, but with much lower complexity.

KW - Broadband wireless communications

KW - Diversity

KW - Eigen-analysis

KW - Multiple receive antennas

KW - Orthogonal frequency-division multiplexing (OFDM)

UR - http://www.scopus.com/inward/record.url?scp=9444288172&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=9444288172&partnerID=8YFLogxK

U2 - 10.1109/TCOMM.2004.836594

DO - 10.1109/TCOMM.2004.836594

M3 - Article

VL - 52

SP - 2019

EP - 2027

JO - IEEE Transactions on Communications

JF - IEEE Transactions on Communications

SN - 0096-1965

IS - 11

ER -