Study On Behavoral Models Of Radio Frequency Power Amplifier

With the rapid development of modern wireless communication, to reduce product design cycles and increase its precision in the radio frequency microwave field, the key device of communication system-power amplifiers system-level simulation needs the higher demand. Especially the development of the microwave circuit simulation and digital predistortion technology, study on behavioral models of RF power amplifier has the important significance to enhance the performance of simulation system.This article is based on the National Natural Sciences Foundation"Study on nonlinearity and behavioral models for broadband RF power amplifier", study on behavioral models of power amplifier adopts the Matlab software and the sampled data of ADS software simulation, the main content include:Introduces power amplifier characteristics and common behavioral models, simulates and validates the complex coefficient polynomial model, the Saleh model, the memory polynomial model, the Volterra model as well as the RBF neural network model.Improves behavioral models based on Volterra series, include: the behavioral model based on the Saleh function simplifies the original Saleh model parameters solution process, and the Volterra-Laguerre behavioral model simplifies Volterra model coefficient and enhances the precision.Based on the BP neural network model, proposes the memory BP neural network and PSO_BP behavioral models enhance model precision. In addition, studies and models theεSVM (support vector machine) forecast behavioral model based on PSO optimizing model parameters.In order to describe the actual RF power amplifier more precisely, models the dynamic behavior model of the power amplifier, simulation indicates that the improved models of PA have the stable precision and good auto-adapted ability. In addition, analyzes that how to confirm the actual performance of behavioral models through the experiment measurement. In the end, introduce the actual construction and application of behavioral models.