Respiration and Heart Rate Detection for Human Dynamic Monitoring Based on Millimeter-wave Radar

Respiratory rate and heart rate are important indicators of the physiological state and psychological changes of the human body. In recent years, millimeter-wave radar-based respiration and heart rate estimation techniques have received much attention due to their non-invasiveness and cost-effectiveness. However, the accuracy of heart rate estimation decreases significantly in dynamic scenarios such as body movements or conversations, which becomes a major challenge for the development of this technology. To address this problem, this paper proposes a millimeter-wave radar-based respiration and heart rate detection method specifically optimized for dynamic scenarios. Specifically, in order to reduce the interference caused by body movement in dynamic environments, we design a novel denoising method that effectively reduces the noise introduced by factors such as body swaying, and thus reconstructs clearer and more accurate millimeter-wave radar detection signals. In the detection phase, we further propose a hybrid network model combining CNN, LSTM and Transformer architectures. The model is capable of accurately identifying key points in millimeter-wave radar time-series signals, which significantly improves the accuracy of respiration and heart rate detection in dynamic scenes. The experimental results show that the proposed method performs well in dynamic environments and provides strong support for the deployment and application of millimeter-wave radar in complex scenarios.