Engineers at the National University of Singapore have developed a wearable sensor capable of tracking fatigue and stress levels based on physiological signals—even during physical activity. The study was published in the journal Nature Sensors.

One of the main problems with existing fitness trackers is a sharp drop in accuracy during movement: muscle interference masks weak signals from the heart and blood pressure. To address this, the researchers created a platform based on a meta-hydrogel—a soft material that adheres to the skin. Nanoparticles within the gel absorb mechanical vibrations, a special electrolyte suppresses high-frequency noise from muscles, and machine learning algorithms further clean the signal. As a result, the signal-to-noise ratio increased from 5.19 to 37.36 dB—meeting clinical standards and surpassing most commercial trackers.

The machine learning system estimates fatigue levels with an accuracy of about 92%. Fatigue and stress are reflected in the functioning of the autonomic nervous system—through changes in heart rate variability, blood pressure, and ECG characteristics.

According to the authors, the technology is especially relevant for people whose jobs require high levels of concentration, such as drivers, air traffic controllers, doctors, and operators of complex machinery. In addition, the sensor can detect breathing, heart sounds, brain waves, and eye movements, opening up broad possibilities for health monitoring.