Hardware designers nowadays need to be fully aware of all high-frequency phenomena that cause signal integrity (SI) issues. Moreover, adverse effects are also induced by further miniaturization, as manufacturing tolerances lead to undesired geometrical and material parameter variability. Therefore, to reach their stringent design specifications, engineers have to rely on efficient modeling methods that accurately capture all high-frequency effects and quantify parameter uncertainty.
State-of-the-art uncertainty quantification techniques, e.g. based on generalized polynomial chaos expansions, rely a considerable amount of (expensive) input data with known variability. In this tutorial, however, novel Machine Learning (ML) based strategies to construct generative models of stochastic link responses are explained. Starting merely from a small set of "training" responses, obtained either from simulations or measurements, ML allows us to construct comprehensive stochastic models from which new response samples can be generated with a distribution as similar as possible to the real data distribution, for use in Monte Carlo-like analyses.
The ML techniques are thoroughly explained and abundantly illustrated using commercial connector and printed circuit board interconnect data. Comparisons on the prediction of frequency- and time-domain responses are discussed, demonstrating the appositeness of ML-based generative modeling for future SI-aware design of high-speed interconnects and circuits.
Dries Vande Ginste received the M.S. and Ph.D. degrees in electrical engineering from Ghent University, Ghent, Belgium, in 2000 and 2005, respectively. In 2004, he was a Visiting Scientist with the Department of Electrical and Computer Engineering, University of Illinois at Urbana-Champaign, Urbana, IL, USA. In 2011, he was a Visiting Professor with the EMC Group, Dipartimento di Elettronica, Politecnico di Torino, Turin, Italy. He is currently an Associate Professor with the Department of Information Technology, Ghent University, and a Guest Professor at imec, Ghent. He has authored or co-authored over 150 papers in international journals and in conference proceedings. His current research interests include computational electromagnetics, electromagnetic compatibility, signal and power integrity, and antenna design. Dr. Vande Ginste was awarded the International Union of Radio Science (URSI) Young Scientist Award at the 2011 URSI General Assembly and Scientific Symposium, the Best Poster Paper Award at the 2012 IEEE Electrical Design of Advanced Packaging and Systems Symposium (EDAPS), the Best Paper Award at the 2013 IEEE Workshop