New techniques for non-linear behavioral modeling of microwave/RF ICs from simulation and nonlinear microwave measurements

This paper compares and contrasts recent nonlinear behavioral modeling techniques designed for microwave and RFIC application which arise in radio and communication systems, and in the design of broad-band nonlinear components used for microwave instrumentation. These techniques include dynamic neural networks and nonlinear time series models in the time-domain, nonlinear describing functions in the frequency domain, and envelope-based methods in mixed time and frequency domains. Approaches to generating these models from both simulation and nonlinear microwave measurements are reviewed.