Research On Several New Issues Of Blind Equalization Based On CM

The transmission characteristics of wireless communication channels have influences on the properties of received signals. The distortion compensation technique is an important approac h to improve the quality of received distortion signals and reliability of communication systems. Channel equalization is an effective measure for distortion compensation, and the third party usually uses the blind equalization technique. Traditional blind equalization skills aimed at linear time invariant channels provided that the transmitted signals are circular. In fact, the properties of signals and channels are complex as well as various. The transmitted signals can not only be circular, but also be noncircular. In some surroundings, the channel is not linear time invariant but nonlinear or time variant. Thus, we start from the characteristics of transmitted signals and channels, and the receiving mode to probe new blind equalization technique to impro ve the performance and practicability of blind equalizer. The paper focus mainly on three issues, blind equalization of complex-valued noncircular signals, blind equalization of nonlinear time invariant channels based on the Hammerstein model and linear time variant channel on space diversity.Firstly, in order to improve the blind equalization performance of complex-valued noncircular signals, such as 8QAM and the rectangular 32 QAM, et al., a new widely linear(WL) blind equalization structure based on constant modulus(CM) criterion is constructed, then the WL constant modulus algorithm(WL-CMA) and WL recursive least mean square constant modulus algorithm(WL-RLS-CMA) are proposed. New algorithms accounted for the full second-order information of the comp lex-valued noncircular signals and make up for deficiency of the conventional blind equalization algorithms. Theoretical inference has proved that new algorithms can obtain better performance than conventional algorithms. Simulation results also show the superiority of the new algorithms.Secondly, in order to improve the performance of nonlinear channel blind equalizers and reduce the complexity, the Hammerstein model is introduced to simulate the nonlinear channel instead of traditional Volterra series model. Then a new nonlinear channel WL blind equalizer based on Wiener model is produced by use of the noncircularity of received signals which are transmitted by nonlinear channels, thus the nonlinear channel WL constant modulus algorithm(NCWL-CMA) and the nonlinear channel WL constant modulus Newton-like algorithm(NCWL-CMA Newton-like) are proposed. Theoretical analysis and simulation results show that the new algorithms can significantly reduce the intersymbol interference and improve the convergence rate.Finally, to expand the application scope of the CM-based linear time variant single input multiple output(SIMO) blind equalization algorithms, an improved CM-based linear time variant SIMO blind equalization algorithm without channel limitation is proposed. Based on deep research of the existing algorithms, the problem that the existing algorithms are restrained by channel limitation has been analyzed. The improved algorithm has demonstrated that output of the SIMO equalizer may have more than one basis frequencies without channel limitation. Thus a new basis frequency estimation algorithm is proposed, and a new updating algorithm of the equalizer coefficients is introduced. Simulation results show that the improved algorithm is unrestrained with channel condition, and has better convergence performance and practicability compared with existing algorithms.