论文部分内容阅读
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。
In this paper, we propose a complex valued minimum error rate non-linear filter for QAM signal equalization in nonlinear channels. The objective function of training algorithm for the lowest bit error rate (BER) criterion for QAM signal is deduced, and the Volterra sequence is used to realize the complex value nonlinear equalizer. In order to enable the nonlinear equalizer to train on-line adaptively and increase the numerical stability of the training algorithm, a sliding window stochastic gradient algorithm is proposed. A large number of simulations show that the performance of the lowest error rate nonlinear equalizer is better than the least mean square error criterion for QAM signal equalization in nonlinear channels.