Research On Digital Predistortion Technology For Power Amplifiers

Power amplifiers play an important role in modern communication systems.However,due to their inherent nonlinearity and memory effects,the transmission signal bandwidth is expanded,resulting in distortion during signal transmission and affecting the quality of the received signal at the receiving end.In order to improve the spectral efficiency of the system,modern communication systems adopt efficient modulation methods,such as orthogonal frequency division multiplexing and quadrature amplitude modulation.These non-constant envelope signals are more susceptible to nonlinear effects,thereby reducing the performance of the communication system.Therefore,solving the nonlinearity problem of power amplifiers has become a hot research topic in wireless communication transmitter systems.Among various linearization techniques,Digital Pre-Distortion(DPD)technology has become the focus of power amplifier linearization technology research due to its digital implementation,strong adaptive ability,and low cost advantages.In this thesis,the behavior model of power amplifier and pre-distortion device,pre-distortion structure and adaptive parameter identification algorithm in digital predistortion technology are systematically analyzed,and the parameter identification algorithm of pre-distortion device is focused on in-depth study.The power amplifiers are divided into two categories according to the presence or absence of memory effect:amplifiers without memory and amplifiers with memory.In order to study the amplifier system adaptability of the pre-distortion device based on various parameter identification algorithms,simulation experimental analysis is conducted in the amplifier system without memory and with memory behavior model respectively.This thesis first conducts a simulation analysis on the pre-distorter based on traditional parameter identification algorithms.To address the limited linearization performance of the pre-distorter based on the traditional Least Mean Squares(LMS)algorithm,two improved variable step-size LMS algorithm-based pre-distorter parameter identification algorithms are proposed in this thesis,and compared with the pre-distorter based on the Normalized Least Mean Square(NLMS)algorithm.Simulation results show that the convergence performance of the pre-distorter based on the proposed algorithms is superior and it can better improve the nonlinearity of the power amplifier system.The pre-distorter evaluation indicators,including the Advisory Committee of Permanent Representatives(ACPR)and the Normalized Mean Square Error(NMSE),are both better than those of the pre-distorter system based on the NLMS algorithm.To further improve the performance of the predistorter based on the improved variable step-size LMS algorithm,this thesis proposes an improved variable step-size LMS algorithm based on Adaptive Chaotic Particle Swarm Optimization(ACPSO)for parameter identification in the predistorter.The simulation results are compared with the predistorter based on the corresponding improved variable step-size LMS algorithm.The simulation results show that in both memoryless and memory behavior model power amplifier systems,the digital predistorter based on the two improved variable step-size LMS algorithms using ACPSO can further improve the non-linear characteristics of the power amplifier system.Compared with the digital predistorter using the corresponding improved variable step-size LMS algorithm,the predistorter based on ACPSO has better convergence performance and better improvement in the non-linearity of the power amplifier system.In terms of the evaluation indicators ACPR and NMSE,the performance of the predistortion system based on ACPSO is better than that of the corresponding improved variable step-size LMS algorithm.It further improves the performance of the predistorter.Moreover,in the memoryless behavior model power amplifier system where the deviation of the predistorter model is large,the performance index of the proposed algorithm is greatly improved compared with other algorithms,indicating that the proposed algorithm also improves the adaptability of the predistorter in different power amplifier systems.