Microwave Large Signal Scattering Function Modeling Based On BP Neural Network

With wide applications of microwave technology, the design of microwave circuit becomes more and more attentive. Whereas, Modern communication tends to larger power level and more sophisticated modulation method, which makes the nonlinear characteristic of components and circuits of the system. Therefore, accurate modeling on large signal nonlinear behavior becomes popular in recent research.S-parameter theory is a linear theory, no longer fits for large-signal analysis. This paper researched a new large-signal analysis method, namely large-signal scattering function. The method takes the influence of harmonic waves into account, so it can be used for the analysis of nonlinear characteristic.A black-box model based on frequency domain measurement is introduced. In this paper, we introduce how to measure power transistors in a real large signal situation. BP Neural Network (BPNN) is used to model on microwave components because it has been proved the most accurate. Finally, we verified the accuracy of this BPNN model.