In the real communication systems, the channel is always unideal. ,which may induce ISI.In order to mitigate ISI, the equalizer is used. The commomly used Equalization algorithms are Least Mean-Square (LMS), Recursive Least-Squares (RLS), Reduced Constellation Algorithm (RCA), Constant Module Algorithm (CMA). The first two algorithms need training sequences, but the third and forth algorithms do not require them, which belonged to blind equalization (BE). Usually, excess mean square error (MSE) and convergence speed are two criterions used to evaluate performance. Step-size parameter is a key factor to decide convergence speed and excess MSE. A fast convergence speed needs large step-size, but a small excess MSE requires small one. It means that these two criterions can not be optimized concurrently when using fixed step-size. For solving this problem, the variable step-size technology is introducted in equalization.Variable step-size equalization algorithm has both the fast convergence speed and small excess MSE.This thesis focuses on the usage of variable step-size technology in equalization. First, a new boundary of step-size in equalization is... |