EM-Based Statistical Analysis and Yield Estimation Using Linear-Input and Neural-Output Space Mapping

Abstract

We propose a method for highly accurate electromagnetics-based statistical analysis and yield calculations around a space mapped nominal solution for RF and microwave circuits. Our method consists of applying a constrained Broyden-based linear input mapping approach to design, followed by an output neural mapping modeling process in which not only the responses but the design parameters and independent variable are used as inputs to the output neural network. We illustrate the accuracy of our technique using a classical synthetic problem