An optimal selection of classic twobox behavioral models for RF power amplifiers

This paper presents a method that aims at selecting the optimal behavioral model and figuring out the fingerprint of the RF power amplifiers. First of all, it introduces the main classification of the behavioral models. Then, regarding Wiener model and Hammerstein model, it proposes an algorithm with a detailed mathematical explanation, which is based on weighted least squares method and obtains the optimal behavioral models successfully. Finally, it presents a numerical simulation for RF power amplifiers in the real world and validates the practicability and effectiveness of the algorithm. In the simulation, firstly the parameters of the behavioral models are computed through the training signals, then the optimal behavioral model is selected by analyzing the errors according to various judgment criteria. The result of numerical simulation reflects the algorithm could obtain the optimal behavioral models of the RF power amplifiers effectively with a relatively small error.