In order to improve the performance of constant modulus algorithm(CMA) in the noise which obeys α-stable distribution, wavelet frequency domain constant modulus blind equalization algorithm based on fractional lower order statistics(WT-FLOSFCMA) is proposed. This proposed algorithm uses fractional lower order statistics to restrain α-stable noise, and its computational loads can be greatly reduced by using FFT technique and the overlapping retention law. In the proposed algorithm, according to the minimum dispersion coefficient criterion, the CMA is optimized, orthogonal wavelet transform is used to improve the convergence rate. The computer simulations in underwater acoustic channels show that the proposed algorithm outperforms than CMA, orthogonal wavelet transform constant modulus algorithm(WT-CMA),and frequency domain blind equalization algorithm based on fractional lower order statistics(FLOSFCMA) in suppressing with inter-symbol interfere, improving convergence rate, and reducing mean square error.
A modified blind equalization technique based on a constant modulus algorithm
