Statistical Design of High Frequency Electronic Circuits Using Space Mapping-Based Neuromodels
Description
Manufacturability-driven designs in a time-to-market development industrial environment demand accurate yield optimization and statistical analysis of the electronic components. Yield optimization requires intensive simulations to cover the entire statistic of possible outcomes of a given manufacturing process. Full wave electromagnetic simulations are usually needed to accurately characterize high frequency electronic circuits. Performing direct yield optimization using accurate full wave electromagnetic simulations for high frequency circuits does not appear feasible. In this paper, an efficient procedure to perform electromagnetics-based yield optimization and statistical analysis of high frequency structures using space mapping-based neuromodels is proposed. Our technique is illustrated by the EM-based statistical analysis and yield optimization of an HTS filter.ITESO, A.C.
Natural Sciences and Engineering Research Council of Canada
Consejo Nacional de Ciencia y Tecnología