Bussgang Blind Equalization Algorithm And Fpga Implementation

As a result of the channel distortion caused by nonideal channel frequency response characteristic, a succession of pulses transmitted through the channel at rates comparable to the bandwidth are smeared to the point that they are no longer distinguishable as well-defined pulses at the receiver, i.e., the intersymbol interference (ISI) occur among a number of adjacent symbols. Channel equalization is used to avoid or mitigate the problem of ISI. Since it is common for the channel characteristics to be unknown, or to change over time, the equalizer is designed to be capable of tracking a slowly time-varying channel response. Classical equalization techniques employ a time slot during which a training signal, known in advanve by the receiver, is transmitted. The training signal is transmitted periodically in time-varying situations. Blind channel equalization, without resorting to a known training sequence, can not only adapt itself to the slowly changing environment, but also save the bandwidth of transmission. There are three major families of blind equalization techniques: the first is the family of stochastic gradient algorithms, which is called Bussgang algorithms; the second is based on the higher-order, or cyclostationary statistics; the third is the maximum likelihood algorithm based on the mutual information. Bussgang algorithms are especially suitable for hardware implementations. Bussgang algorithm provides a uniform framework for decision-directed algorithm, Sato algorithm, Godard algorithm, Benveniste-Goursat algorithm, and Stop-and-Go algorithm. Constant modulus algorithm, a special case of Godard algorithm, is the most studied and implemented blind equalization technique. In this dissertation, the design principles of receiver and the generation of intersymbol interference are presented. Both the zero-forcing equalization and the MMSE equalization are discussed. The Bussgang blind channel equalization algorithms, including the decision-directed algorithm, Sato algorithm, Godard algorithm, Benveniste-Goursat algorithm, and Stop-and-Go algorithm, are presented in great detail. The analysis results are supported by the computer simulations of the CMA algorithm and DDLMS-Godard algorithm. An implementation of blind equalization algorithm on FPGAs is also presented. The DLMS algorithm is useful for pipeline hardware implementations. The implementation details including the finite...