A Study of Fast Converging Blind Equalizers


  1. Kozminchuk, B.
  2. Wang, S.
Corporate Authors
Defence Research Establishment Ottawa, Ottawa ONT (CAN);Telexis Corp, Ottawa ONT (CAN)
Two blind identification and equalization algorithms based on second order statistics are investigated. These algorithms have been reported in the literature as having fast convergence, and have been suggested for equalizing signals propagating in time-varying channels. However, two significant difficulties with these algorithms were discovered. First, it was found that they are sensitive to slight inaccuracies in the signal subspace dimension. Standard methods used to estimate the signal subspace dimension provide values which are in the vicinity of the desired one, but which may require adjustment by the user to obtain good blind equalization results. The second difficulty is that the two algorithms enhance noise for low oversampling ratios to such an extent that the transmitted symbol constellation may not be identifiable, even when the SNR is as high as 40 or 50 dB and even when the dimension of the signal subspace is correctly estimated. To overcome this limitation, the oversampling ratio must be increased, which has the desirable effects of decreasing the noise enhancement and improving the convergence rate of the blind equalizewr for QAM and M-PSK signals at SNRs as low as 25 to 30 dB. TRUNCATED
Modulation recognition;Synchronization;Equalizer;Blind equalizer;CMA (Constant Modulus Algorithm)
Report Number
DREO-TM-1999-097 — Technical Memorandum
Date of publication
01 Nov 1999
Number of Pages
Hardcopy;Document Image stored on Optical Disk

Permanent link

Document 1 of 1

Date modified: