Abstract: Aiming at the problem of predistortion for memory nonlinear amplifier in predistortion technology, this paper analyzes the structure of digital predistortion and recognization algorithm of common predistortion, improves classical Least Square/Singular Value Decomposition (LS/SVD) algorithm. Improved LS/SVD algorithm can obtain better performance by less resource. Simulation results show the proposed algorithm can realize fast and effective linearization of memory nonlinear amplifier, and improve its performance.

Key words: digital predistortion, Hammerstein model, Least Square(LS), Singular Value Decomposition(SVD)

摘要：

针对预失真技术中存在记忆非线性放大器预失真的问题，分析数字预失真器的结构和常用预失真器的识别算法，对经典最小二 乘/奇异值分解(LS/SVD)算法进行改进，以较少资源获得较高性能。仿真结果表明，改进的LS/SVD算法能实现记忆非线性放大器的快速、高效线性化，提高记忆非线性放大器的性能。

关键词: 数字预失真, Hammerstein模型, 最小二乘, 奇异值分解

CLC Number: 

• TP391.9