摘要: 讨论一种基于正交递归最小二乘法(QR-RLS)的功率放大器行为模型。该模型采用Gives旋转提高QR-RLS算法的运算效率,能够提升数字预失真模型的系数更新速度,更快地实现数字预失真模型的收敛。测试一个44 dBm的两载波WIMAX功率放大器,并基于测试数据建立动态数字预失真模型。分析结果表明,该模型能校正宽带功率放大器的非线性特性,并快速实时地获得模型参数。
关键词:
行为模型,
Volterra级数,
记忆效应,
功率放大器,
递归最小二乘法
Abstract: In this paper, a power amplifiers behavioral model based on QR-RLS algorithm is introduced. In this model, the Gives circumgyration, which can efficiently enhances the calculation efficiency of QR-RLS algorithms, is used to improve the update speed of Digital Predistortion(DPD) model coefficients, and in turn, fasten the convergence speed of the DPD model. It measures a 44 dBm two-carrier WIMAX Power Amplifier(PA) and builds the dynamic DPD model based on input and output data obtained form the test. Performance analysis of dynamic model shows the model can not only describe and correct the nonlinear characteristic of wide-band power amplifier efficiently, but also obtain the value of model coefficients in faster speed and more dynamic method.
Key words:
behavior model,
Volterra series,
memory effect,
Power Amplifier(PA),
Recursive Least Square(RLS) method
中图分类号:
王敏, 王联国, 刘成忠. 基于QR-RLS算法的预失真模型[J]. 计算机工程, 2011, 37(14): 280-281.
WANG Min, WANG Lian-Guo, LIU Cheng-Zhong. Predistortion Model Based on QR-RLS Algorithm[J]. Computer Engineering, 2011, 37(14): 280-281.