Semi-blind methods for FIR multichannel estimation

The purpose of semi-blind equalization is to exploit the information used by blind methods as well as the information coming from known symbols. Semi{blind techniques robustify the blind and training problems, and o er better performance than these two methods. We rst present identi ability conditions and perfor- mance bounds for semi{blind estimation. Semi{blind methods are able to estimate any channel, even when the position of the known symbols in the burst is arbi- trary. Performance bounds for semi{blind multichannel estimation are provided through the analysis of Cramer-Rao bounds (CRBs) and a comparison of semi{ blind techniques with blind and training sequence based techniques is presented. Three categories of semi{blind methods are considered. Optimal semi{blind meth- ods take into account all blind information and all the information coming from the known symbols, even if not grouped. Mainly Maximum{Likelihood (ML) methods will be considered. Suboptimal semi{blind solutions are also investigated when the known symbols are grouped in a training sequence: the suboptimal semi{blind cri- teria appear as a linear combination of a training sequence criterion and a blind ML criterion. Thirdly, we present methods that combine a given blind criterion with a training sequence based criterion.