|
With the development of electronic technology , microwave circuits has been widely used in various fields of communications , radar , electronic countermeasures , as well as instrumentation . How accurate design and analysis of microwave circuits become very critical . The traditional design and analysis methods are established on the basis of the weakly nonlinear , for strongly nonlinear microwave circuit is no longer applicable . Therefore , it is an urgent need to put forward a new theory to solve this problem . This paper describes the traditional microwave nonlinear characterization methods , and Agilent signal model , small signal S-parameters on the basis of these methods to expand , and then proposed a new microwave nonlinear characterization methods nonlinear scattering function . Moreover, we also study its physical significance and nature , build a the microwave nonlinear scattering function testing system , to extract the nonlinear scattering function of the semiconductor device measurement data . Neural network method based on black-box model , in theory, can be a good approximation of the nonlinear relationship . Therefore , this paper adopts BP improved algorithm combined with wavelet neural network to model nonlinear scattering function model of microwave semiconductor devices , and get better results , and nonlinear scattering function test system to extract data design microwave power amplifier verify the accuracy of the theory .
|