Research And Implementation Of Digital Predistortion In LTE-A System

With the rapid development of mobile communications,generations of communication technology have been proposed and widely used.The LTE-A system can use carrier aggregation technology to extend the baseband signal to a bandwidth of up to 100 MHz,while the increase in the baseband bandwidth will lead to more serious non-linear distortion of the amplifier.This thesis uses LTE-A system as the background to analyze the key technology of broadband predistortion.It analyzes the characteristics of LTE-A system power amplifier's feature extraction platform,LTE-A system digital predistortion model and broadband low sampling technology.The main work and innovations of this thesis are summarized as following points:1.This thesis analyzes the impact of the imbalance between transmit and receive channels in the feature extraction platform of the broadband power amplifier on the power amplifier modeling.Aiming at solving the frequency imbalance of IQ in the feature extraction platform of the broadband power amplifier,this thesis proposes a time-domain based broadband IQ imbalance compensation algorithm by combining the existing compensation technology.A digital predistortion experiment is performed on a continuous amplifier which was driven by 20 MHz single-carrier.Compared with no compensation,the ACLR of the wideband channel compensation model proposed in this thesis was improved by 6dB and 8dB,respectively.2.This thesis analyzes the characteristics of the behavior model of the power amplifier under the wideband signal excitation.This thesis first studies the existing models and analyzes the distribution of various basic functions and the complexity of model,then propose a DDMP model with better balance between complexity and modeling accuracy,which has better modeling accuracy and lower complexity comparing with the widely used existing mode.In the VSG-VSA instrument platform,the digital pre-distortion experiment is verified by driving a single carrier 20 MHz baseband signal.Under the same conditions,the digital predistortion effect of the proposed DDMP model is 3 dB higher than the DDRV model's,and the coefficient of the model is 30% less than 1-MDDRV's.3.This thesis analyzes the problem of limited feedback bandwidth of digital predistortion in wideband communication systems,and proposes a differential sampling PA estimation model and alignment algorithm.The feedback loop of the LTE-A system of digital predistortion is analyzed.An improved PA model estimation method for differential sampling is proposed through the study of several existing methods of low sampling.The PA model is constructed by the low sampling feedback signal and the fully sampled input signal.And then the fully sampled output signal is estimated.The improvement of the proposed method in this thesis is mainly to solve the unstable modeling accuracy of the PA model with high complexity and low sampling rate under different memory depth and nonlinear order.Through simulation and experimentation,the low sampling multiples are set to be 10,50,and 100 times respectively.Compared with the existing low sampling technology,both the complexity of the alignment algorithm and the recovery of the full sampling output signal are greatly improved.In particular,when the low sampling multiple is 50,the running time of the alignment algorithm proposed in this thesis is 10% of the existing algorithm,and the modified PA model estimation method for differential sampling maintains a high accuracy modeling area(NMSE<-30 dB)above 90% for different memory depths and nonlinear orders.