Vital signs monitoring based on millimeter-wave radar using permutation entropy optimized VMD

Millimeter-wave radar, with high resolution and superior penetration, shows considerable potential for vital signs monitoring, particularly in detecting respiratory and heartbeat signals. This paper explores its application in this domain. Due to the weak nature of heartbeat signals, the Variational Mode Decomposition (VMD) method, known for its excellent capability in handling multimodal signals, has been widely applied to extract heartbeat signals. Nevertheless, Challenges persist in accurately extracting respiratory and heartbeat signals from the Intrinsic Mode Functions (IMFs) of VMD decomposition and in identifying optimal number of modes and penalty factor. Vital signals have a certain periodicity, and permutation entropy can be used to separate respiratory heartbeat signals as a periodic characterization of the time series. A novel VMD method optimized through permutation entropy (PE-VMD) is proposed in this paper and designed to enhance the accuracy and stability of vital sign monitoring. Initially, a permutation entropy minimization strategy is proposed in the paper to refine VMD parameters for effective vital signal decomposition. Respiratory and heartbeat signals are then accurately isolated and extracted by analyzing the correlation between the IMF components from VMD decomposition and signals processed through spectral filtering. The paper measured respiration and heart rates from extracted signal frequencies, compared them with other optimized VMD methods, and validated the results using electrocardiogram(ECG) data. The experimental results demonstrate that the PE-VMD method achieves respiration and heart rate measurement accuracies of <strong>90.91%</strong> and <strong>92.83%</strong>, respectively, surpassing alternative methods and confirming its high precision and robustness in vital signal extraction.