| High spectral efficiency digital modulation technique was used to improve the spectrum efficiency, but it produced a non-constant envelope signal which required a higher linearity of the power amplifier. Typically, the method which improved the linearity of the power amplifier was reducing its output power, so that the power amplifier was operating at the linear region, but this would decrease the efficiency and had a waste of power resources. Linear technology can improve power efficiency at the condition of not reducing the linearity of the power amplifier. In the research of linearity technology, we often give an appropriate model of the power amplifier and a simulation of the system to prove the validity of an algorithm, design a linearization program. So, in a design of system level, it is very important to model the power amplifier. Among the various linearization techniques, digital pre-distortion technology is a highly efficient linear technology, which uses digital circuits to correct the distortion produced by the nonlinearity of power amplifier, it has a high efficiency, stable performance and other advantage and will be a trend of linearization technology development. In this dissertation, it begins with modeling power amplifier and linearization technology. The main researches are as follows:1. After a study of the non-linearity of power amplifier, analyze the nonlinear characteristic of power amplifier and important indicators for linearity of power amplifier, describe the impact on the communication system of nonlinearity and analyze the relationship between linearity, spectral efficiency and power efficiency. The causes which generate power amplifier memory effect are analyzed. Give details about several important factors which cause the memory effect.2. Describe the modeling behavior theory for power amplifier at the beginning, and analyze the structure and parameter estimation methods of various power amplifier models, including Saleh model without memory, Volterra series model with memory, Wiener model and Hammerstein model. Describe the choice of the power amplifier model and how to verify the Validity of amplifier model. Take power amplifier MRF21125 for example, build amplifier circuit simulation with ADS2006A, uses the simulation data instead of amplifier test data, verify the validity of the model of the amplifier by the curve fitting method.3. Research on the structure of polynomial model of the power amplifier. The parameter estimation of traditional polynomial model exist a numerical instability problem. To avoid the numerical instability problem, it presents a model with orthogonal polynomial basis functions, and proves that when the order of polynomial increases, the mean square error between the behavior model and circuit model would decrease, it can increase the accuracy of modeling amplifier, and can be extended to a polynomial model with memory effect.4. A study on the methods about adaptive digital pre-distortion algorithm. Describe the fundamental principles and structures of several linear technologies including the feedback method, feed-forward method, and digital pre-distortion method. Based on indirect learning structure, proposes a Recursive Prediction Error (RPE) algorithm of digital pre-distortion, which has a faster convergence than the Recursive Least Square (RLS) algorithm. Design a program to implement digital pre-distortion and analyze the performance of the program and its main components in the circuit functions. Give the whole board and has a test, the test indicates that the program has some performance of improving the power amplifier linearity. |
PDF Full Text Request