SIMULATION OF NOVEL QRD BASED ALGORITHM FOR BLIND EQUALIZATION

Abstract

As real world communication channels are stressed with higher data rates, intersymbol' interference (IS I) becomes a dominant limiting factor. One way to curb this effect is to use an equalizer in the receiver. Equalization for digital communications constitutes a very particular blind deconvolution problem in that the received signal is cyclostationary. Oversampling of the cyclostationary received signal leads to a stationary vector-valued signal. Oversampling leads to a fractionally spaced channel model and equalizer. Blind fractionally spaced equalizers reduce intersymbol interference using second-order statistics without the need for training sequences. A new blind equalization method is proposed that exploits the cyclostationarity of oversampled communication signals to achieve equalization of possibly nonminimum phase (multipath) channels. Most of the eigenstructure-based blind channel identification and equalization algorithms with second order statistics need singular value decomposition or cigen value decomposition of the correlation matrix of the received. signal. This dissertation addresses a novel QR factorization algorithm based on linear prediction approach. It performs the QR factorization of the received signal directly without calculating the correlation matrix. Another advantage is that it estimates the transmitted data directly, without resorting to channel identification stage. Performance of the proposed methods and comparisons with existing methods are shown. Also shown the performance of the method wuth respect to channel undcrmodeling/overmodeling.