The use of artificial neural networks in nonlinear microwave devices and circuits modeling: An application to telecommunication system design (invited article)

The design of telecommunication systems is a hierarchical process involving use of a large set of simulation tools and relying on appropriate modeling of system elements to obtain get-it-right-the-first-time fabrication. This paper investigates in detail possible application of neural networks to modeling of large-signal hard-nonlinear behavior of power transistors for circuit design purpose, and modeling of nonlinear circuits such as power amplifiers for system design purpose. The problem of finding a good model is discussed through solutions offered by neural networks, with particular interest in wavelet networks trained by BFGS algorithms. Finally, experimental results are reported.