Research On Adaptive Digital Predistortion Technology Suitable For New DLA

With the development of mobile communication system and the strengthening of the awareness of green environmental protection,as the indispensable and power-consuming component of mobile communication systems,RF power amplifiers(PA)have become increasingly contradictory between efficiency and linearity.In recent years,due to the rapid development of digital signal processing technology and large-scale integrated circuit technology,digital pre-distortion technology based on digital circuit has been widely studied by scholars at home and abroad.This paper mainly studies the adaptive digital pre-distortion technology adapted to the new direct learning architecture.Firstly,the nonlinear characteristics of PA,the classical model and the principle of digital pre-distortion are introduced.Secondly,to deal with the complexity of the traditional direct learning architecture,the new direct learning architecture based on the instantaneous complex gain of PA is put forward.In this architecture,the instantaneous complex gain is used to replace the PA model,and the output error of the pre-distorter is used to replace the output error of the PA to update the coefficient of the pre-distorter.The simulation results show that the new architecture can greatly reduce the computational complexity on the premise of ensuring the pre-distortion performance.Thirdly,aiming at the problem that the classical adaptive algorithm is prone to algorithm error fluctuation at the peak of the input signal,an amplitude segmentation method is proposed to update the pre-distortion coefficients separately at the peak and non-peak of the input signal,which can effectively reduce the algorithm error fluctuation.Fourthly,in view of the adaptive algorithm can't adjust in time to result in the increase of algorithm error,the pre-distortion coefficient of the real-time update of the adaptive algorithm is regarded as the real-time update coefficient(RUC)and the concept of Memory Correction Coefficient(MCC)is put forward.The RUC-MCC binding algorithm and RUC-MCC fusion algorithm are proposed to solve the problem of algorithm instability.Finally,the experimental results show that the new direct learning structure based on instantaneous complex gain can reduce the computational complexity with guaranteeing the performance of pre-distortion and the new pre-distorter coefficient extraction algorithms have better performance because of their good stability.