Digital Predistortion Research In Wireless Communication

In modern communication system, in the face of the booming data volume andincreasingly scarce spectrum resources, people usually use digital transmissiontechnology with high frequency spectrum utilization and modulation mode with highefficiency. But these transmission technologies and modulation modes generally havea constant envelope, wide band and high peak-to-average ratio, signal can producenonlinear distortion when they pass through power amplifier. But this kind ofdistortion can’t get rid of by filtering, so power amplifier linearization technologyhave wide application and development. Digital predistortion technology is currentlyrecognized the most promising power amplifier linearization technique, it isconvenient to realize by using digital signal processing device.This dissertation firstly discusses the principle of nonlinear distortion, the memoryeffect as well as several linearization indicators of power amplifier. Then, severalmain power amplifier linearization techniques are compared, here select predistortiontechnology as research. Modeling and analysis is the foundation of power amplifierlinearization, power amplifier is divided into memoryless model and memory model.Establish a suitable model according to the specific power amplifier can be helpfulfor the research, so as to achieve the best predistortion effect. Memory polynomialpower amplifier model is actually the most used, and it is a compromise modelbetween memoryless power amplifier model and Volterra series model. This paperadopts the look-up table predistortion scheme with indirect learning structure, usingMatlab language to describe the whole system on Matlab platform, so as to realize thefunction of predistortion and adaptive algorithm, here proposes RLS_DCD (RLSusing Dichotomous Coordinate Descent) look-up table update algorithm. Also, onSystem Generator tool, choose needed modules and establish the digital predistortionsystem. Simulation and analysis results have shown the correctness of the system andthe validity of the algorithm, which improve ACPR index more than20dB both forthe memoryless power amplifier and memory power amplifier.