Research On Digital Predistortion Technology For 5G Broadband Wireless Communication

With the continuous development of modern wireless communication systems towards highspeed,large capacity and ultra wideband,communication signals with many characteristics such as multi carrier,multi-level,ultra-high bandwidth and peak to average ratio are also widely used,which also puts forward higher requirements for the linearization of transmission signals of communication systems.The power amplifier is the main module for introducing nonlinearity into the communication system.In order to ensure the linearity of the transmitted signal,many nonlinear compensation technologies have been developed.Digital predistortion(DPD)has become the most commonly used nonlinear compensation technology for power amplifiers due to its high flexibility,high accuracy and low cost.With the development of 5G communication system,how to reduce the sampling rate of the DPD feedback loop Analog to Digital Converter(ADC)and the cost and power consumption of the DPD module has become a problem to be solved in order to meet the requirements of the 5G communication system’s ultra bandwidth signal and the miniaturization of the base station.At the same time,with the close combination of Multiple Input Multiple Output(MIMO)technology and beam forming technology in 5G communication system,considering that it is difficult to set DPD feedback loop for MIMO antenna array spacing,and MIMO system corresponds to a large number of RF links and power amplifiers,the traditional single RF link DPD is no longer applicable,so the nonlinear compensation problem of MIMO system gradually changes from power amplifier DPD to beam DPD.Therefore,this thesis focuses on the DPD technology of single RF link and nonlinear compensation in MIMO scenarios.This thesis first introduces the relevant theoretical basis of power amplifier digital predistortion in communication systems,in which the measurement index of nonlinear correlation,the basic types of existing power amplifier nonlinear models and the corresponding power amplifier nonlinearity under MIMO beamforming architecture are described in detail.Then,the basic learning structure of power amplifier digital predistortion is summarized.Finally,the basic principles and algorithm flow of common model parameter identification methods are introduced.Secondly,this thesis studies the power amplifier nonlinear compensation technology of single RF link,and proposes a power amplifier DPD method based on indirect learning structure and single channel down sampling feedback,aiming to reduce the ADC sampling rate in the DPD feedback loop and reduce the cost and complexity of engineering implementation.The ADC of the feedback loop of the DPD scheme can collect the same direction components of the output signal of the power amplifier as feedback data at a rate lower than the Nyquist sampling rate of the input signal of the power amplifier,and complete the digital predistortion process of the signal at a high rate through an indirect learning structure.Then,the key technology of DPD,namely time alignment,is introduced,and a time alignment scheme in the case of down sampling is proposed.Then,the effectiveness of the proposed DPD scheme for power amplifier nonlinear compensation is verified through simulation.The signal power spectral density,AM/AM and AM/PM characteristic curves,EVM,constellation diagram and bit error rate under the DPD scheme are analyzed in detail,and the performance is compared with that of the traditional two-way full sampling feedback DPD scheme.The simulation results show that the proposed DPD method for power amplifier nonlinear compensation is equivalent to that of the traditional full sampling feedback DPD scheme,And it greatly reduces the sampling rate requirement of the feedback loop ADC and the cost and complexity of engineering implementation.Finally,the proposed DPD scheme is verified by the physical system.Finally,this thesis studies the nonlinear compensation technology of beamforming in MIMO system,and proposes a beam oriented digital predistortion(BO-DPD)technology for MIMO system based on single channel over the air(OTA)and downsampling feedback in a single beam scenario,On the basis of reducing the cost and complexity of the DPD system,the nonlinear compensation of the main beam signal of the antenna array at the transmitter is completed.The BO-DPD method is to set a dedicated DPD near-field receiving antenna and conduct down sampling in the feedback link,solve the model coefficients of each power amplifier of the antenna array at the transmitting end through multi angle signal testing,and then estimate the user’s main beam far-field receiving signal to complete the DPD compensation process of the main beam signal.Then,the DPD solution process is derived,and the direct solution method to solve the parameters of each power amplifier model of the antenna array and the indirect solution method to reduce the computational complexity are introduced.Then,by comparing the power spectral density of the actual received signal and the estimated signal of the main beam through simulation,the accuracy of the feedforward models of the antenna array power amplifiers of the BO-DPD method is verified,and the signal power spectral density,normalized mean square error(NMSE),AM/AM and AM/PM characteristic curves,and the adjacent channel power ratio(ACPR)in the BO-DPD process are analyzed in detail,The simulation results show that the BO-DPD has obvious nonlinear compensation performance for the main beam signal in the single beam scene of the MIMO system.