Human Body Movement Estimation Based on Spatial and Temporal Separating Features for Multiple People

Continuous and non-invasive vital sign monitoring using radar sensors is crucial for healthcare applications. However, accurately separating multiple human movements in home sleeping scenarios remains challenging. This paper proposes a novel method for human body movement estimation method based on spatial and temporal separating features for multiple people. Specifically, our method begins with target capture using signal processing techniques such as fast fourier transform (FFT), clutter suppression, and capon beamforming. Then, we perform target separation and adaptive target region estimation by calculating density spectra and identifying clusters based on peak detection. We refine the location estimates of detected clusters and stabilize their positions over time using a correction formula. Subsequently, we estimate human body movements by applying directional beamforming to isolate signals for each individual, performing FFT over time to obtain distance-angle-Doppler information, and integrating frequency components to summarize movement patterns. Experimental results show that the proposed method is effective in both single-target action estimation and dual-target scenarios.