Unified pre-processing steps reducing the PCB design space to enable ML applications for signal and power integrity analysis

Citation
29th IEEE Workshop on Signal and Power Integrity, SPI 2025
Contribution to Conference
29th IEEE Workshop on Signal and Power Integrity, SPI 2025  
Publisher DOI
10.1109/SPI64682.2025.11014326
Scopus ID
2-s2.0-105007665997
Publisher
IEEE
ISBN of container
979-8-3315-2061-8
979-8-3315-2062-5
Machine Learning (ML) applications need large quantities of data spanning the complete Printed Circuit Board (PCB) design space. This design space needs to be reduced to the subset of relevant PCB configurations, i.e. those yielding network parameters which are different from each other. For this purpose a preliminary design space analysis is carried out and pre-processing steps are presented to unify the application of ML in the Signal and Power Integrity domains and pave the way to a standardized PCB representation format.
Subjects
Database
Machine Learning
Power Integrity
Printed Circuit Board
Signal Integrity
DDC Class
621.3: Electrical Engineering, Electronic Engineering