| With 5G wireless communications on its way,under the requirements of high spectrum efficiency and high data rate,the linearization methods on radio frequency PAs have been switching from analog ones to digital ones in wireless communications.Digital predistortion system has become the main stream upon linearization methods due to its excellent linearization performances,simple implementation as well as highly adaptive characteristic.Besides,digital pre-distortion system is capable of linearizing different kinds of PAs because it only deals with baseband signals.This article puts the focus on a realistic communication system with closed-loop working situation and aims to build up a flexible structure,highly adaptive digital pre-distortion system basing on an FPGA platform.In this article,a novel on-line closed-loop digital pre-distortion system structure is proposed.This structure is able to tune its pre-distorter parameters following the PAs' nonlinear characteristics.A new amplitude normalization procedure is proposed in order to get rid of the divergence of model parameter estimation algorithm,which is not stable in the traditional off-line closed-loop structures and is likely to go divergence because of amplitude estimation errors.Under a simulation of variable load impedance situation,this structure has been proved to have outstanding linearization performances and stabilities,far superior to the traditional off-line structure.In addition,a closed-loop digital pre-distortion system basing on FPGA platform is designed and implemented in this article.The design is flexible to easily modifying its model structure and desired adaptive algorithm before compilation.The complex root-leastsquare algorithm is implemented in the design,which then is equipped with high accuracy calculation similar to PC.Under a realistic closed-loop test,the implementation achieved about 24 dB ACPR compression,which is approximately the maximum linearity of the whole analog circuits. |
PDF Full Text Request