A Low-Complexity Blind Equalization Algorithm for 4096-QAM Systems.
Bo Wu,Nan Zhang,Hang Zhang,Feng-Kui Gong
DOI: https://doi.org/10.1109/apcc.2018.8633561
2018-01-01
Abstract:A simplified blind equalization algorithm, which can substantially reduce the complexity of hardware implementation of neighborhood-assisted symbol-based decision (N-SBD) algorithm, is proposed for 4096-QAM systems. Two main strategies are adopted in order to achieve our goals. On the one hand, the step function is used to approximate the weighting factor in the conventional N-SBD algorithm for avoiding the complicated exponential calculations. On the other hand, by utilizing the equidistant characteristics for the adjacent symbol coordinates, we derive the formula of the error function again, thus significantly reducing the number of multipliers in implementation. The analysis further shows that, for 4096-QAM signals, the proposed algorithm greatly reduces the implementation complexity at the cost of reducing the convergence speed slightly, whereas it can maintain the mean square error (MSE) performance of the steady-state residual error when compared with the conventional N-SBD algorithm. Simulation results also indicate that both the convergence speed and the steady-state residual error of the proposed algorithm are superior to other blind equalization algorithms, such as the multi-modulus algorithm (MMA) and constellation matched error-based MMA (CME-MMA).