Wavelet Transform Based Blind Equalizer Design And Algorithm Simulation

At present, the report on the research of channel blind equalization processing, which combined the WT (Wavelet transform) with blind equalization technology, is very few at home and abroad. But it is a very significant topic to research on channel blind equalization technology by means of wavelet theory and making full use of its superiority, and it is also an advanced topic in modern signal processing field. Aiming at the disadvantage of CMA (Constant Modulus Algorithm), the blind equalization theory, algorithm and simulation based on wavelet transform are studied deep in this dissertation. The main contributions are as follows:1. Aiming at the contradiction of CMA with a constant step-size between the convergence rate and the residual error, the variable step-size theory is applied to CMA and several variable step-size CMAs, including REVCMA( the exponent function of Residual Error based Variable step-size CMA), RDVCMA (Rayleigh Distribution based variable step-size CMA) and MCMA(a new Modified CMA), are put forward. Simulation results indicate that these proposed algorithms can overcome the contradiction of CMA between the convergence rate and the residual error.2. On the basis of researching on WT-LMS (Wavelet Transform based Least Mean Square algorithm),a new WT-CMA(Wavelet Transform based CMA) is proposed. The new algorithm integrate WT with adaptive blind equalization, the convergence rate of the new algorithm can be improved by using the good decorrelation ability of WT. Simulation results showe its convergence rate is faster and its computation complexity is smaller.3. MWT-CMA (momentum term and orthogonal wavelet transform based blind CMA equalization Algorithm) and VWT-CMA (Variable step-size and orthogonal wavelet transform based CMA blind equalization Algorithm) are proposed. In the proposed algorithms, the WT-CMA is integrated with variable step-size theory and momentum term respectively, the convergence rate can be improved by using the good convergence performance of variable step-size and momentum term. Simulation results testify the above effectiveness of these algorithms.4. Aiming at M-QAM signal, WT-MMA (orthogonal Wavelet Transform based Multi-Modulus blind equalization Algorithm) is proposed. The proposed algorithm retained the good decorrelation character of WT and the little residual error character of Multi-Modulus Algorithm (MMA), so the convergence rate is faster and the residual error is smaller. As not all of the output signals data are used in WT-MMA, so the convergence rate is slow. In order to improve the convergence rate, WT-SDDMA (orthogonal Wavelet Transform based Sign Decision Dual-Modulus blind equalization Algorithm) is proposed further. The algorithm can switch between WT-CMA and WT-MMA in terms of sign decision. So its is best. Simulation results are accord with theory analysis.