摘要
在数字预失真过程中,针对传统的最小均方差(LMS)自适应算法收敛速度较慢和稳态误差较大的缺点,通过分析各种改进的变步长LMS算法的优势和劣势,提出了一种组合曲线的归一化变步长LMS数字预失真算法。仿真结果表明,新算法在保证预失真效果的前提下,算法性能得到极大提升。
In the process of digital pre-distortion,to deal with the shortcoming of the slower convergence speed and larger steady-state error in typical LMS adaptive algorithm,by analyzing the strengths and weaknesses of various improved Least Mean Square（LMS） algorithms,a normalized variable step-size LMS digital pre-distortion algorithm with combination of curve is proposed. The simulation results show that the proposed algorithm can improve performance extremely and ensure the accuracy of pre-distortion.
出处
《微型机与应用》
2017年第20期80-83,91,共5页
Microcomputer & Its Applications
基金
四川省教育厅科研项目(15ZB0458)
四川工商学院科研项目(2016ZYB17X)
关键词
预失真
最小均方差
变步长
归一化
pre-distortion
Least Mean Square(LMS)
variable step-size
normalization
