The Research On The Adaptive Algorithm Of Baseband Predistortion Of Power Amplifier Linearization

Nowadays, when the third generation communication system is being built for broad business application, and the prototype of the fourth generation mobile communication system has been put forward to, all their base stations depend on the microwave devices with high power and high linearization, the key of which is the power amplifier linearization technology. It is fundamental to raise the whole system's capability under the condition that the power amplifier linearization and its efficiency is contradict with each other. In this dissertation, based on the comprehensive analysis of all kinds of linearization technologies introduced before, and on the adequate study of the nonlinear characteristic of power amplifier, it chose the adaptive algorithm of baseband predistortion of power amplifier linearization as the target of the dissertation. The predistorter creates a complementary nonlinearity, which ideally is the exact inverse of the PA nonlinearity to a scaled input, and leads to the result that the the overall predistirter-PA chain is approximately linear. It is highly efficient, stable, wide-band and easy to make adaptive, and becomes the most potential and the most effective technology for RF PA linearization.The main studying of dissertation is about basic theories and basic methods, main research is as follow:1 analyze the nonlinearization of power amplifier and the influence to the system of communication, introduce the major characteristic of non-linear power amplifier, analyze several commonly-used mathematical model of power amplifier, present an overview of the common linearization techniques in existence, and compare their advantages and disadvantages. Then give a specific introduction of the principle of predistortion and adaptation, which sets up the base of the research.2 Propose a modified digital adaptive polynomial predistorting technique. firstly, adopt the polynomial predistortion in nonlinearization to fit for the HPA inverse model, and use RLS adaptation algorithm to adjust the coefficient of polynomial, which lead to the polynomial match the inverse model of HPA exactly. It is easy to realize, effective and stable, but the computation is huge. Based on the principle of slowly-adaptive predistortion, and the fact, which is verified by Matlab simulation, the IMD 3 and the IMD 5 is mailly affected by the odd-order nonlinear terms of the polynomial function and less affected by the even-order nonlinear terms, A modified digital adaptive polynomial predistorting technique is proposed. The basic principle of this polynomial predistortion method is to approximate the target function with two simpler functions-the function of the odd-order nonlinear terms and the even-order nonlinear terms instead of the conventional polynomial function. So we can compute the RLS algorithm with much less complexity. The simulation results show that it can decrease computation greatly, meanwhile the IMD3 is depressed with 34dB, and IMD5 with 58dB.