Neural-network approaches to electromagnetic-based modeling of passive components and their applications to high-frequency and high-speed nonlinear circuit optimization

摘要: In this paper, artificial neural-network approaches to electromagnetic (EM)-based modeling in both frequency and time domains their applications nonlinear circuit optimization are presented. Through accurate fast EM-based neural models of passive components, we enable consideration EM effects high-frequency high-speed computer-aided design, including component's geometrical/physical parameters as variables. Formulations for standard frequency-domain approach, recent time-domain approach based on state-space concept, described. A new combining existing knowledge the form equivalent circuits (ECs), with equations (SSEs) networks (NNs), called EC-SSE-NN, is proposed. The EC-SSE-NN allow behaviors components interact active devices, facilitate domain. An automatic mechanism data generation, which can lead efficient training Demonstration examples a three-stage amplifier, multilayer printed board, geometrical/physical-oriented power-plane effects, interconnect embedded terminations buffers domain