Un nouvel égaliseur: L'égaliseur a double échantillonnage
Abstract
The paper describes a method for realizing global optimization of an adaptive nonrecursive equalizer in which the sampling moment is taken into account. The nonlinearity of such an equalizer precludes the immediate use of classical optimization algorithms based on the steepest descent. Sampling at the input is done at a frequency twice that of the data emission rate. This enables an interpolation of the input samples corresponding to any sampling moment. Thus, the input samples become linear quantities with respect to the parameter describing the sampling moment. At this point, then, it is possible to proceed along the classical lines of optimization using the steepest descent method. The resulting equalizer is similar to classical equalizers, except that it treats simultaneously two channels in phase opposition.
 Publication:

Annals of Telecommunications
 Pub Date:
 September 1975
 DOI:
 10.1007/BF03003168
 Bibcode:
 1975AnTel..30..331M
 Keywords:

 Adaptive Filters;
 Channels (Data Transmission);
 Data Sampling;
 Equalizers (Circuits);
 Optimal Control;
 Automatic Control;
 Binary Digits;
 Convergence;
 Error Analysis;
 Intersymbolic Interference;
 Linear Systems;
 Nonlinear Systems;
 Self Adaptive Control Systems;
 Signal Distortion;
 Steepest Descent Method;
 Communications and Radar