Semi-blind equalization at the symbol rate with application to OFDM

  • Authors:
  • Surendra Prasad;Anubala S. Varikat

  • Affiliations:
  • Department of Electrical Engineering, Indian Institute of Technology, New Delhi 110 016, India;Department of Electrical Engineering, Indian Institute of Technology, New Delhi 110 016, India

  • Venue:
  • Signal Processing
  • Year:
  • 2003

Quantified Score

Hi-index 0.08

Visualization

Abstract

Blind equalization algorithms for non-minimum-phase channels are generally based on oversampling the output of the system. In this paper, we consider a symbol-rate sampled scheme that requires the periodic transmission of a few zeros within the data sequence. Sampling at the symbol rate relaxes some of the identifiability conditions, and ensures that the additive noise is uncorrelated. We develop a least-squares approach for the blind equalization of such a system, that works well with very short observation windows. The price paid for eliminating the oversampling is the reduced throughput. We derive the necessary and sufficient conditions for the identifiability of the system. Our simulations show that channels with near-common zeros can be equalized better using the proposed scheme, as compared to existing schemes. Also, the proposed method is seen to perform much better than the recently proposed methods (IEEE Trans. Signal Process. 47 (1999) 2007) in terms of bit error rate, for short observation windows, and requires much less stringent identifiability conditions. However, the computational complexity of the method is high. The method can be applied directly to the blind equalization of orthogonal frequency division multiplexing systems that use a zero prefix instead of a cyclic prefix.