By combining AI algorithms with microcontroller based edge processing, the system claims to deliver real time and accurate health monitoring without increasing system complexity.
HOLTEK’S Edge AI, Smarter at the Edge
As healthcare devices become smaller and more portable, there is growing demand for on device intelligence that can deliver accurate health measurements without relying on cloud processing. Wearable and card sized medical devices, in particular, require low power operation, stable performance and real time data processing at the edge.
Addressing this requirement, Holtek has expanded its HT32 microcontroller lineup with Edge AI focused solutions designed for measurement, wearable and AI computing applications. The new offerings target intelligent sensing and health monitoring use cases where compact form factor and energy efficiency are critical.
One highlighted application is a card type blood pressure meter that supports non-invasive monitoring of blood pressure, heart rate, and blood oxygen saturation. The system uses an on-device neural network combining convolutional neural networks and back propagation neural networks to filter noise and predict readings. According to the company, the measurement results comply with AAMI standards.
The solution is built around an HT32 series microcontroller and applies pulse transit time technology along with AI based calibration. Transfer learning is used to adapt measurements to different body types, while continuous optimization improves accuracy with repeated use.
Key features include:
- On device CNN plus BPNN algorithm for blood pressure prediction
- Real time measurement with continuous self optimization
- Lightweight AI computation suitable for microcontrollers
- Low power operation for portable health devices
- Integrated monitoring of blood pressure, heart rate and oxygen saturation
According to the company, the approach demonstrates how Edge AI can enable real time health analytics directly on compact hardware.
