复值的MBER非线性均衡器用于QAM信号

Complex-Valued Minimum Bit Error Rate Nonlinear Equalizers for QAM Signals

本文提出一种复值的最低误码率非线性滤波器用于非线性信道中QAM信号的均衡。推导了针对QAM信号的最低误码率准则训练算法的目标函数,并用Volterra序列来实现复值的非线性均衡器。为使非线性均衡器能在线自适应训练并增加训练算法的数值稳定性,提出一种滑窗随机梯度算法。大量仿真表明,对于非线性信道中QAM信号的均衡,最低误码率非线性均衡器的性能优于最小均方误差准则。

A complex-valued minimum bit error rate （MBER） nonlinear equalizer is proposed for equalization of quadrature-amplitude modulation （QAM） in nonlinear channels. The MBER objective function for QAM is derived and complex-valued nonlinear equalizers are realized via Volterra series. In order to train the equalizers online and improve robustness of training, a sliding-window stochastic gradient algorithm is proposed. Extensive simulations show that the performance of MBER equalizers is superior to that of minimum mean square error equalizers for QAM in nonlinear channels.