Radar-based recognition of respiration, heart rate, sleep, and posture

A research team from National Taiwan University collaborates with ShangMao Intelligence Technology to develop a millimeter-wave radar detection system. The system utilizes feature engineering and AI technology to detect the subject's current respiratory rate, heart rate, sleep stage, and posture. In the later stage of development, the algorithm will be transplanted into Nvidia's edge devices to complete the project development. The developed system uses the phase information detected by millimeter-wave radar and integrates multiple algorithms to extract features and predict results, achieving excellent performance in detection. The proposed method can currently perform stably in most situations, but outliers still need to be addressed through data set expansion and algorithm improvement. To improve accuracy and the algorithm's recognition capability, the team is actively developing new architectures and expects to complete the development with an accurate and efficient solution.