PAM-4 Behavioral Modeling using Machine Learning via Laguerre-Volterra Expansion

Xinying Wang, Thong Nguyen, Jose E. Schutt-Aine

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

In this paper, we demonstrate a PAM-4 IBIS-AMI model obtained from machine learning for time domain simulation in commercial simulators. More specifically, we report a Laguerre-Volterra-expanded Feed-Forward neural Network (LVFFN) with only one hidden layer and 10 neurons to model the 28Gb/s PAM-4 high speed link buffer. the LVFFN model reduces the model size and improves the computational efficiency dramatically compared to the Volterra series model and other transitional artificial neural network models. The LVFFN PAM-4 model can be implemented in IBIS-AMI, an industrial standard for high speed link modeling. Eye-diagram analysis is conducted on LVFFN PAM-4 IBIS-AMI model in commercial simulators. Our results show that our model is faster to simulate 1 million bits and is easier to implement compared to the standard PAM-4 IBIS-AMI model.

Original languageEnglish (US)
Title of host publication2020 IEEE 11th Latin American Symposium on Circuits and Systems, LASCAS 2020
PublisherInstitute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)9781728134277
DOIs
StatePublished - Feb 2020
Event11th IEEE Latin American Symposium on Circuits and Systems, LASCAS 2020 - San Jose, Costa Rica
Duration: Feb 25 2020Feb 28 2020

Publication series

Name2020 IEEE 11th Latin American Symposium on Circuits and Systems, LASCAS 2020

Conference

Conference11th IEEE Latin American Symposium on Circuits and Systems, LASCAS 2020
CountryCosta Rica
CitySan Jose
Period2/25/202/28/20

Keywords

  • Behavior Modeling
  • Eye Diagram
  • IBIS-AMI
  • Laguerre
  • PAM-4

ASJC Scopus subject areas

  • Hardware and Architecture
  • Energy Engineering and Power Technology
  • Electrical and Electronic Engineering
  • Instrumentation

Fingerprint Dive into the research topics of 'PAM-4 Behavioral Modeling using Machine Learning via Laguerre-Volterra Expansion'. Together they form a unique fingerprint.

Cite this