Wearable embedded edge-computing device for real-time classifying tremors in Parkinson’s disease

초록

This study presents a wearable embedded edge-computing device for real-time classifying hand tremors in Parkinson’s disease, aiming at demonstrating the real-time classification based on supervised deep learning algorithm embedded on resource-constraint tiny micro controller (tinyML). Parkinson’s disease is typically characterized by uncontrollable, rhythmic shaking, most often beginning unilaterally in the fingers or hands. Thus, it is difficult to characterize or sense by using the only sensors. In this study, a lightweight 1-D convolutional neural network-based classifier was embedded on tiny edge-computing device. The proposed embedded edge-computing device can be integrated with active tremor stabilizer with actuator (e.g., Gyroscopic moment actuator), and anti-tremor controllers. © 2026 SPIE.

키워드