论文部分内容阅读
本文针对包络跟踪功率放大器预失真系统,提出一种新的分段分数阶记忆多项式模型(PFMP),并基于该模型设计数字预失真器。其采用矢量阈值分解技术将输入复包络信号分解成几个子信号,并以分数阶记忆多项式为每个子信号构建数字预失真的子函数。这样能使每个不同的区域里的ETPA内的非线性特性都能精确地被表达,且提高了带外抑制性能。仿真结果表明,对于Volterra模型的功率放大器,新模型与传统的记忆多项式模型相比,NMSE降低了约3d B,ACPR降低了约2.3d B。
In this paper, a novel piecewise fractional memory polynomial model (PFMP) is proposed for the envelope tracking power amplifier predistortion system, and a digital predistorter is designed based on the model. It uses vector thresholding to decompose the input complex envelop signal into several sub-signals, and constructs a digital predistortion sub-function for each sub-signal with a fractional memory polynomial. This enables the non-linearities within the ETPA in each of the different areas to be accurately expressed and improves out-of-band rejection. The simulation results show that for the Volterra model power amplifier, the NMSE is reduced by about 3dB and the ACPR is reduced by about 2.3dB compared with the traditional memory polynomial model.