arrWNN: Arrhythmia-detecting Weightless Neural Network FlexIC
Pillai, V.
; Miranda, I.
; Musale, T.
; Jadhao, M.
; Susskind, Z.
; Bacellar, A.
; Ozer, E.
; Lhostis, M.
; Lima, P.
; Dutra, D.
; John, E.
; Breternitz Jr, M.
; França, F.
; John, L.
arrWNN: Arrhythmia-detecting Weightless Neural Network FlexIC, Proc IEEE International Flexible Electronics Technology Conference IEEE IFECT, Bologna, Italy, Vol. , pp. - , September, 2024.
Digital Object Identifier:
Abstract
This paper proposes a technique for incorporating machine learning into a wearable medical patch by combining two key technologies: weightless neural networks (WNNs), known for their efficiency and low hardware needs, and Flexible Integrated Circuits (FlexICs) - ultra low-cost circuits on flexible substrates. We designed a special WNN called 'arrWNN' for detecting arrhythmia events from ECG data. Then, we optimised arrWNN for area efficiency, and fabricated it using Pragmatic's FlexIC technology. Our wafer-level test and measurement results found a number of fully functional arrWNNs.
