基于小波变换变步长LMS自适应信道均衡的新方法

Adaptive Channel Equalization Using Variable Step Size LMS Algorithm Based on Wavelet Transform

结合变换域最小均方(LMS)和变步长LMS算法的优势,提出了一种基于小波变换的变步长LMS自适应均衡方法。该方法中步长调整函数采用了改进的Sigmoid函数,该函数具有简单且误差信号接近零时变化缓慢的特点。并且,在训练模式、判决引导模式以及混合模式下,将提出方法和传统均衡方法进行了仿真比较。结果表明,所提出的方法比传统的线性LMS算法、变步长LMS以及小波变换LMS收敛更快、性能更优。

By combining the advantages of transform domain least mean square （LMS） and variable step size LMS algorithms,an adaptive linear equalization LMS algorithm using variable step size based on wavelet transform domain is proposed, where the step size is adjusted as an improved Sigmoid function. This function is sim- ple and varies slowly as the error signal approaches zero. Moreover, the proposed method is compared with the traditional channel equalization methods in the training mode, the decision-directed mode and the hybrid mode, respectively. Simulations show that the proposed method may offer higher convergence rate and better equalization performance than the conventional LMS method, the variable step size LMS method and the LMS method based on wavelet transform domain.