基于直接学习法的一种新预失真器

为了克服宽带信号经过记忆放大器的非线性失真,针对有记忆非线性功放的多项式模型,提出了一种新的基于直接学习法的自适应算法.该算法采用无记忆预失真器的级联扩展,具有横向滤波器结构,与记忆多项式有相似的线性化效果.并且针对信号噪声对自适应算法的扰动和收敛速度慢等缺点,采用归一化LMS算法加以改进.在非线性功放的记忆多项式模型下,通过宽带信号验证了基于直接学习法的记忆型预失真器算法的有效性.

Novel Adaptive Nonlinear Predistorter Based on the Direct Learning Algorithm

A novel adaptive direct learning algorithm for power amplifier with memory effects is proposed to compensate the nonlinear distortion introduced by HPAs.The cascade extension of the memoryless predistorter and the structure of finite-impulse-response filter significantly can get the close performs compare to memory polynomial method.The Normalize LMS algorithm is proposed to the disturb of signal noise and convergence speed problems.Under power amplifier’s memory polynomial model,simulation result show that the predistorter has the excellent performance when the wideband signal pass it.