.png){kind=icon}
.png){kind=icon}
Return
Ultralow-Power Floating-Gate InGaZnO Content-Addressable Memory for Wearable Electrocardiogram Anomaly Detection
.png){kind=icon}
OA
AI
H
S
Y
J
S
DOI：10.1002/aisy.70474.png)
.png){kind=icon}
.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Ultralow-power operation is essential for continuous wearable electrocardiogram (ECG) monitoring under a limited power budget. This work proposes an in-memory anomaly detection system based on floating-gate (FG) InGaZnO (IGZO) content-addressable memories (CAMs), operating in the subthreshold regime. By engineering the area-based capacitive coupling ratio inherent in the FG structure, the subthreshold slope (SS) and subthreshold operating window are precisely controlled at the layout level without modifying materials or process conditions. The proposed CAM intrinsically performs exponential distance computation in the subthreshold regime, thereby improving detection performance. This computation is directly mapped to a linear match-line voltage response, which significantly enhances the sensing margin and enables robust anomaly detection even with low-resolution analog-to-digital converters. At the system level, on-chip in-memory inference eliminates continuous data streaming and high-resolution digitization, resulting in an estimated 50.5× reduction in on-chip energy per heartbeat and a substantial extension of battery lifetime. These results establish FG IGZO CAMs operating in the subthreshold regime with high SS as an effective hardware platform for energy-efficient ECG anomaly detection and broader edge-AI applications.
Keywords:
anomaly detection
content-addressable memories
electrocardiograms
InGaZnO
subthreshold operation
subthreshold slope
wearable devices
AI Summary
Key information extracted from the uploaded paper, including a brief overview, abstract, background, key highlights, visual analysis, and future outlook.
Journal
IF:
6.1
Papers:
1.9K
Citations:
8.4K
Organization
Cited Papers