An IoT-enabled embedded wearable system using Edge AI achieved 99.8% accuracy and 70% lower latency compared to standard models for continuous cardiovascular health monitoring.
The proposed Edge AI-empowered wearable system provides highly accurate and low-latency continuous cardiovascular monitoring on resource-constrained devices.
The rising incidence of cardiovascular diseases necessitates the development of real-time, intelligent, and privacy-preserving health monitoring systems. Traditional wearable solutions often rely on cloud processing, which compromises latency, energy efficiency, and user privacy. To address these challenges, this study proposes an advanced IoT-enabled embedded wearable system empowered by Edge AI for continuous cardiovascular health monitoring. The system integrates a novel hybrid deep learning model combining 1D Depthwise Separable Convolution (1D-DSC) and Temporal Attention-Gated Recurrent Unit (TA-GRU) to capture both spatial and temporal dependencies in ECG, PPG, and heart rate signals. This study is implemented on an ARM Cortex-M−based microcontroller using TensorFlow Lite. The model achieves exceptional performance with 99.8 % accuracy, 99.62 % precision, 99.54 % recall, and an AUROC of 0.997. The model size is compressed from 22 KB to 11 KB using post-training quantization (PTQ), reducing inference time from 15 ms to just 7 ms without significant loss in accuracy (a drop of only 0.25 %). A real-time alert system visualizes abnormal conditions through LED color codes and haptic feedback, ensuring immediate user awareness. Compared to state-of-the-art models like CNN, Bi-LSTM, and 1D CNN-LSTM, the proposed method exhibits superior accuracy and 70 % lower latency. This work demonstrates a compelling solution for deploying deep learning models on resource-constrained wearables, enabling energy-efficient, secure, and continuous cardiovascular monitoring in real-world scenarios.
Rani et al. (Thu,) conducted a other in Cardiovascular diseases. IoT-enabled embedded wearable system with Edge AI (1D-DSC and TA-GRU) vs. CNN, Bi-LSTM, and 1D CNN-LSTM models was evaluated on Model accuracy for cardiovascular health monitoring. An IoT-enabled embedded wearable system using Edge AI achieved 99.8% accuracy and 70% lower latency compared to standard models for continuous cardiovascular health monitoring.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: