The Design And Realization Of Digital Predistortion Linearizers For RF Power Amplifier

In wireless communications, the radio frequency (RF) power amplifiers (PA) may result in the increase of inter modulation (IM) distortion and adjacent channel interference for several modulation approaches due to its inherent nonlinear characteristic. Therefore, the performance of PA will severely affect the quality and efficiency of the whole communication system. A typical approach is to use linearization technique to proofread the non-linear characteristic, which will increase the efficiency of transmission, reduce the energy consumption, and improve the adjacent channel power radio (ACPR). With the development of digital processing technology and digital processor, digital predistortion becomes one of the most effective approaches among all linearization techniques. The objective of this work is to develop an efficient digital predistortion method.In this thesis, the parameters which are commonly used to evaluate an RF power amplifier are introduced firstly before a study on the nonlinear characteristics of PA. The mechanism to generate the nonlinear distortion and the nonlinear performance of PA are investigated. Two PA models, the memoryless model and memory model, are considered, and the indirect learning method is used in the research.A FPGA-based indirect learning PA scheme is suggested. Based on the Volterra polynomial, a tracking model to combat with the memory nonlinear distortion in the proposed scheme is presented, and the LMS algorithm is used. The proposed scheme exhibits the efficiency in computation and the simplicity in hardware implementation. The efficiency is validated by using MATLAB-based simulations. The scheme is finally realized by designing a related hardware system as well as the FPGA software system. The experiment results of the system demonstrates the effectiveness of the proposed scheme.