Aportación a la linealización de amplificadores de potencia mediante la teoría de la hiperestabilidad

Abstract

This thesis deals with the use of the parallel model reference adaptive control to the RF power amplifier's linearization problem, by exploiting the use of the passivity (hyperstability) property in the design of the corresponding adaptation law. As a result of our investigation, we obtain a linearizer structure with comparable or inferior complexity to that of other linearization techniques currently in use, and which is capable of reducing the AM/AM and AM/PM distortion with a comparable or superior performance. Although the concept of passivity has been studied in the field of the Control Theory, up to now it had not been carried out a detailed study for its application to the linearization of power amplifiers (PA) at radio frequencies (RF), considering both the phase and the amplitude distortions. Our contribution is the study of the application of the Adaptive Control Theory and the concept of the Passivity in Cartesian form to the solution of the RF power amplifier linearization problem, matter up to day unpublished, and the proposal of a novel linearizer with high performance in comparison with other well consolidated linearization techniques. This thesis is divided in 7 Chapters. In the first Chapter it is presented a general perspective of the problem, and both the global and the specific objectives of the investigation are settled down. In Chapter 2 we characterize the power amplifier input and output signals formally, we describe the digital modulation formats that concern to the problem and we treat the matter of the PA nonlinearities modelling. In Chapter 3 we present the problem antecedents: we describe in concise way some different linearization techniques that have been developed so far: feedforward, predistortion, feedback, EE&R, LINC and CALLUM. In the chapter 4 we develop the theoretical concepts relative to analog dissipative systems, and to the subset of passive systems, lossless (hyperstables) and strictly passive (asymptotically hyperstables). We define, also, the problem of the model reference adaptive control systems. In Chapter 5 we apply the theoretical concepts presented in Chapter 4 and we design a power amplifier linearizer. In this chapter we prove different adaptation laws combinations for linearizing a PA that exhibits amplitude distortion and, moreover, phase distortion. Finally, we propose several structures for the physical implementation of the linearizer. Among these structures one is analyzed in detail. In Chapter 6, using as benchmark the experimental data of a power amplifier for applications at 28 GHz, we test our linearizer and we contrast it with other linearization schemes (Cartesian feedback and feedforward), at the same scenario, and we check the performance of the designed linearizer. In Chapter 7, finally, we present our conclusions and we offer some alternatives for future investigations in this area.