Dual-Mode Blind Equalization Algorithm for High-Order QAM

A novel dual-mode blind equalization algorithm for high-order quadrature amplitude modulation (QAM) is presented. The proposed algorithm consists of the generalized mult-mudulus algorithm (GMMA) for initial convergence during blind startup, and the decision-directed least mean square algorithm (DDLMS) which is used to enhance the performance. The two operation modes are automatic switched according to the equalizer output error level. Combined with step-size parameter variation online in each mode, the blind equalization solution can converge rapidly with low residual tap fluctuation noise. The simulation results show that the proposed techniques converge quickly and reduce the intersymbol interference (ISI) caused by the propagation channel to -43 dB, with about 10 dB improvement contrast to the GMMA. On the other hand, the phase rotation introduced by channel is corrected simultaneously.