Contactless Sleep Quality Monitoring Exploiting Radar Signal

Sleep quality monitoring is vital for managing health risks associated with sleep disorders. Traditional methods, such as polysomnography, are invasive and can disrupt natural sleep patterns. This paper presents a novel, contactless approach to sleep monitoring using ultra-wideband (UWB) radar, providing a non-intrusive solution that detects key indicators of abnormal sleep, including periodic limb movements, frequent body position changes, and prolonged static position. Exploiting deep neural networks, specifically the VGG 16 model, our method achieves 98 % accuracy in classifying sleep quality features, shows its robustness for reliable analysis. The radar's high sensitivity to body movement enables monitoring without the need for wearable sensors, making it a practical alternative for clinical and home applications to a scalable, accurate and comfortable sleep monitoring.