A team led by NRI Institute of Technology introduces MyWear, a wearable T-shirt embedded with physiological sensors and machine learning models, notably SVM, to monitor heart rate variability and detect stress levels with up to 98% accuracy for improved cardiovascular and stress management.
Key points
MyWear integrates ECG sensors into a wearable T-shirt to capture continuous heart rate variability data.
Support Vector Machine classifier achieves 98% stress detection accuracy by optimizing hyperplane separation of HRV features.
Signal preprocessing and motion-artifact filtering enable reliable feature extraction for six machine learning models in real-time monitoring.
Why it matters:
High-accuracy real-time stress monitoring wearable could transform preventive healthcare by enabling continuous stress and cardiovascular risk assessment outside clinical settings.
Q&A
What is heart rate variability?
How does MyWear reduce motion artifacts?
Why use multiple machine learning models?
How is data privacy ensured?
Read full article
Academy
Heart Rate Variability (HRV)
Heart rate variability (HRV) describes the variation in time intervals between consecutive heartbeats, arising from the interplay of the sympathetic and parasympathetic branches of the autonomic nervous system. This variation reflects how well the body adjusts to physical and emotional stress, recovers after exercise, and maintains internal balance.
HRV analysis relies on measuring these beat-to-beat intervals, often through electrocardiogram (ECG) or photoplethysmography (PPG) sensors. Key time-domain metrics include the standard deviation of normal-to-normal intervals (SDNN) and the root mean square of successive differences (RMSSD). Frequency-domain metrics analyze power distribution across low-frequency (0.04–0.15 Hz) and high-frequency (0.15–0.40 Hz) bands, corresponding to sympathetic and parasympathetic activity, respectively.
Health applications of HRV span stress monitoring, athletic performance optimization, and cardiovascular risk assessment. Higher HRV usually indicates a healthy, adaptable heart, while persistently low HRV can signal stress, fatigue, or underlying disease. Researchers and clinicians use HRV trends to personalize wellness plans, guide training recovery protocols, and detect early signs of autonomic dysfunction.
- Time-domain measures: SDNN, RMSSD, NN50 count
- Frequency-domain measures: LF, HF, LF/HF ratio
- Non-linear measures: Poincaré plot analysis
Wearable Sensors for Health Monitoring
Wearable health sensors integrate into garments, watches, or patches to capture physiological data outside clinical settings. Common sensors include ECG electrodes, PPG optical sensors, accelerometers, and temperature gauges. Machine learning algorithms process the raw signals to detect anomalies, estimate stress levels, and provide personalized feedback.
Key steps in wearable monitoring:
- Data acquisition: Sensors embedded in textiles or devices record physiological signals continuously.
- Signal preprocessing: Filtering algorithms remove motion artifacts, electrical noise, and baseline wander.
- Feature extraction: Time-domain, frequency-domain, and non-linear features quantify the signals for analysis.
- Classification and prediction: Machine learning models such as SVM, KNN, and neural networks classify conditions like stress levels or heart anomalies.
- Feedback and visualization: Mobile apps and web dashboards present real-time metrics, alerts, and longitudinal trends for user engagement.
Wearable sensor technology continues to evolve with advancements in flexible electronics, low-power wireless communication, and on-device AI. These innovations make continuous health monitoring more accessible, enabling preventive care, early detection of health issues, and improved self-management of chronic conditions, which collectively support healthier aging and enhanced quality of life.
