Uncertainty Quantification and Machine Learning Techniques Help Generating Digital Twins of Electronic Systems

C3PS Seminar Series by Flavio Canavero Ph.D. 

Flavio Canavero, Ph.D.  
Department of Electronics and Telecommunications 
Politecnico di Torino
Torino, Italy


The accurate prediction and optimization of the performance of complex electronic systems require parametric modeling and statistical analysis of their behavior, for which several techniques have been developed in the last decades. Well-known examples are represented by Monte Carlo (MC), parametric macromodeling, polynomial chaos (PC), worst-case approaches and –more recently–  machine learning techniques.

Digital twins technology implies a pairing of the virtual and physical worlds for a process, product or service. This is achieved through analysis of data and monitoring of systems to prevent problems before they even occur, develop new opportunities and even plan for the future by using simulations.

The above-cited techniques are essential for reaching the mentioned goals. However, none of these approaches provide an ultimate solution for the problem at hand, since they perform differently, e.g., with respect to the number of parameters and the amount of their variability. Hence, a systematic study of various modeling approaches developed in the past for different applications is worth to be conducted in order to clearly identify the merits, limitations and suitability of the methodologies toward the goal of producing predictive models of real objects.

This seminar modestly intends to illustrate some work of the research group lead by the presenter, aimed at a better understanding of the advantages and limitations of several simulation techniques w.r.t. the reliable replication of the behavior of electronic systems. The feasibility and strengths of the advocated methods are demonstrated based on benchmarks and on the statistical assessment of realistic structures employed in digital systems.

About the Presenter

Flavio G. Canavero received his electronic engineering Master degree from Politecnico (Technical University) of Torino, Italy, in 1977 and the Ph.D. degree from the School of Geophysical Sciences of Georgia Institute of Technology, Atlanta, USA, in 1986. He is currently a Professor of circuit theory with the Department of Electronics and Telecommunications, Politecnico di Torino. His research interests include signal integrity and EMC design issues, interconnect modeling, black-box characterization of digital integrated circuits, EMI and statistics for EMC. Dr. Canavero is an IEEE and URSI Fellow; he received several industrial and IEEE Awards, including the prestigious Richard R. Stoddard Award for Outstanding Performance, which is the EMC Society's highest technical award, and the Honored Member Award of the EMC Society. He has been the Editor-in-Chief for the IEEE TRANSACTIONS ON ELECTROMAGNETIC COMPATIBILITY, the V.P. for Communication Services of the EMC Society, and the Chair of the URSI Commission E. Finally, Dr. Canavero has served in several positions (VP, Department Head, Dean of Graduate School…) within the Governing Bodies of his University.