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

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 (g
_m
) states in a 28nm HKMG FeFET technology [1], (ii) 30× tunable range in g
_m
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 . false alarm rate for two patients, (iv) ultra-low energy of 47 fJ/MAC with 1,000× improvement in area compared to alternative mixed-signal MAC.