Efficient OFDM symbol timing estimator using power difference measurements

A. Al-Dweik, S. Younis, A. Hazmi, C. Tsimenidis, B. Sharif

Research output: Contribution to journalArticlepeer-review

25 Scopus citations

Abstract

This paper presents an efficient blind symbol timing estimation scheme for orthogonal frequency-division multiplexing (OFDM) systems with constant modulus constellation. The proposed technique is designed to estimate symbol timing offsets by minimizing the power difference between subcarriers with similar indices over two consecutive OFDM symbols based on the assumption that the channel slowly changes over time. The proposed power difference estimator (PDE) is totally blind because it requires no prior information about the channel or the transmitted data. Monte Carlo simulation is used to assess the PDE performance in terms of the probability of correct timing estimate P lock-in. Moreover, we propose a new performance metric denoted as the deviation from safe region (DSR). Simulation results have demonstrated that the PDE performs well in severe frequency-selective fading channels and outperforms the other considered estimators. The complexity of the PDE can be significantly reduced by incorporating a low-cost estimator to provide initial coarse timing information. The proposed PDE is realized using feedforward and early-late gate (ELG) configurations. The new PDE-ELG does not suffer from the self-noise problem inherent in other ELG estimators reported in the literature.

Original languageBritish English
Article number6092507
Pages (from-to)509-520
Number of pages12
JournalIEEE Transactions on Vehicular Technology
Volume61
Issue number2
DOIs
StatePublished - Feb 2012

Keywords

  • Delay
  • Early-late gate (ELG)
  • Orthogonal frequency- division multiplexing (OFDM)
  • Synchronization
  • Timing estimation
  • Timing offset
  • Wireless channel

Fingerprint

Dive into the research topics of 'Efficient OFDM symbol timing estimator using power difference measurements'. Together they form a unique fingerprint.

Cite this