Energy-Efficient Edge Inference on Multi-Channel Streaming Data in 28nm HKMG FeFET Technology

S. Dutta, W. Chakraborty, J. Gomez, K. Ni, S. Joshi, S. Datta

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

5 Scopus citations

Abstract

We present a system implementing extremely energy-efficient inference on multi-channel biomedical-sensor data. We leverage Ferroelectric FET (FeFET) to perform classification directly on analog sensor signals. We demonstrate: (i) voltage-controlled multi-domain ferroelectric polarization switching to obtain 8 distinct transconductance (gm) states in a 28nm HKMG FeFET technology [1], (ii) 30x tunable range in gm over the bandwidth of interest, (iii) successful implementation of artifact removal, feature extraction and classification for seizure detection from CHB-MIT EEG dataset with 98.46% accuracy and < 0.375/hr. false alarm rate for two patients, (iv) ultra-low energy of 47 fJ/MAC with 1,000x improvement in area compared to alternative mixed-signal MAC.

Original languageEnglish
Title of host publication2019 Symposium on VLSI Technology, VLSI Technology 2019 - Digest of Technical Papers
PublisherInstitute of Electrical and Electronics Engineers Inc.
PagesT38-T39
ISBN (Electronic)9784863487178
DOIs
StatePublished - Jun 2019
Externally publishedYes
Event39th Symposium on VLSI Technology, VLSI Technology 2019 - Kyoto, Japan
Duration: 9 Jun 201914 Jun 2019

Publication series

NameDigest of Technical Papers - Symposium on VLSI Technology
Volume2019-June
ISSN (Print)0743-1562

Conference

Conference39th Symposium on VLSI Technology, VLSI Technology 2019
Country/TerritoryJapan
CityKyoto
Period9/06/1914/06/19

Bibliographical note

Publisher Copyright:
© 2019 The Japan Society of Applied Physics.

Fingerprint

Dive into the research topics of 'Energy-Efficient Edge Inference on Multi-Channel Streaming Data in 28nm HKMG FeFET Technology'. Together they form a unique fingerprint.

Cite this