Research On Predistortion Of Power Amplifier Under Nyquist Undersampling

Driven by high-bandwidth and multi-band 5G communication.RF power amplifier is the core device.Under the excitation of peak-to-average ratio(PAPR)signal,its nonlinearity will worsen the bit error rate of the system and increase energy consumption,thus affecting the communication quality of the whole system.So power amplifier linearization technology has attracted much attention.In this paper,the sparse coefficients of the power amplifier behavior model and the predistortion feedback loop down-sampling technology are studied.This paper is based on the application of under-sampling compression reconstruction theory.Combining with the behavior model of power amplifier and predistortion technology,the innovative points are put forward.The research of power amplifier linearization technology for high-bandwidth communication system is of far-reaching significance.In linearization technology,the accuracy and computational complexity of power amplifier behavior modeling still need to be improved.Based on the sparsity of the core function of Volterra series behavior model,this paper proposes an adaptive sparsity estimation reconstruction algorithm ASP.The identification of sparse systems is equivalent to the problem of signal reconstruction,and a subset of kernel functions is obtained to reduce the complexity of the model.The algorithm dynamically adjusts the sparsity threshold according to the signal characteristics to improve the estimation accuracy,and sets the energy difference threshold to adjust the step size to improve the sparsity estimation speed.Finally,the subspace tracking algorithm is used to restore the signal.The simulation analysis shows that the running time of ASP is 0.01 seconds with that of SP.Signal distortion increased from 50 dB to 250 dB.Compared with MP and GMP models,the precision of trimmed amplifier models is improved by10.7 dB and 3.9 dB.The model coefficients were reduced by 25% and 57.65% respectively.The algorithm improves the reconstruction accuracy while reducing the operation time,and obtains a low complexity power amplifier behavior model.Aiming at the problem of high sampling rate of predistortion feedback loop of dual-frequency power amplifier.An adaptive sparse dual-frequency predistortion structure based on modulated broadband converter is proposed in this paper.The structure consists of memory effect compensator and under-sampled reconstructed feedback loop.The compensator adopts a simplified piecewise linear function model.Undersampling reconstruction refers to the sampling of all-blind-point modulated broadband converter in the predistortion feedback loop,and thenreconstructing the missing fifth-order and higher-order intermodulation signals using adaptive sparse algorithm.It makes the minimum mean square solution of coefficient weight approximate to the optimum,reduces the acquisition error and improves the linearization effect.The simulation results show that,while improving the stability of the system,the NMSE display is2-3dB higher than that of 2D-MP,2D-DDR and 2D-CPWL.ACPR is 15 dBc higher than that without predistortion.It is of great significance to reduce the predistortion sampling rate and improve the linearity of power amplifier.The paper has 51 figures,4 tables and 85 references.