Font Size: a A A

Study For Blind Channel Identification And Equalization Algorithm

Posted on:2006-05-13Degree:MasterType:Thesis
Country:ChinaCandidate:W B ZhengFull Text:PDF
GTID:2168360155465618Subject:Signal and Information Processing
Abstract/Summary:PDF Full Text Request
Blind identification and equalization has been received considerable attention recently in communication and signal processing, the main work of this dissertation is on this topic. At the first, the theoretic basis is expatiated. Then we detailedly discuss how to design criteria for blind deconvolution of non-minimum phase systems. And then we study the time domain approaches for blind identification and equalization, which are based on second-order statistics. Oversampling the received signals will product the cyclostationarity of communication signals, therefore, an equivalent SIMO (Single Input Multiple Output) system model of SISO (Single Input Single Output) system is proposed. Basing on the equivalent SIMO system model, we developed a new recursive algorithm for blind channel identification and equalization of possibly non-minimum phase channels using only second-order statistics. Firstly, the algorithm is analyzed without noise. Then it is expanded to the noise environment with the subspace method. Traditional methods of blind equalization contain lots of SVD, which leads to high complexity of procedure and large computation. Algorithm II proposed in this paper changes the problem of channel parameter estimation into eigenvector solution with relative to correlation matrix by configuring a series of matrixes. So that, the recursive algorithm is given which can obtain all the eigenvectors of channel vague dimension matrix one by one. Further, the analytic solution of blind channel is provided. Simulations show that, compared with other blind equalization algorithms that are not recursive, the estimation performance of this algorithm keeps unchanged. However, this algorithm effectively simplifies the procedure and lowers the complexity of computation and has high practicality.
Keywords/Search Tags:blind identification and equalization, second-order statistics, cyclostationary, recursion
PDF Full Text Request
Related items