Reconstruction of Human Physiological Signal in Partition Wall Based on UWB Signal Time-Frequency Graph

Through-wall physiological parameter monitoring has significant application potential in non-intrusive health surveillance for individuals living alone and in post-disaster rescue operations. To address challenges such as multipath effects and clutter interference, this paper proposes a method for through-wall physiological signal reconstruction based on ultra-wideband signal time-frequency spectrogram. The method effectively mitigates multipath effects and improves the signal-to-noise ratio using Singular Value Decom-position (SVD) and the Least Squares method. It further employs Ensemble Empirical Mode Decomposition (EEMD) and time-frequency spectrogram for signal decomposition and reconstruction, accurately estimating both respiration and heart rates. Ablation experiments demonstrate that this method enhances monitoring accuracy stability and improves signal reliability.