Optimized IMF Extraction Method for Accurate Signal Processing in Bio-Radar Sensors

An optimization algorithm is proposed to select the mode decomposition level of intrinsic mode functions (IMFs) that contain vital signals in bio-radar sensors. The phase spectrum of each IMF is analyzed to determine if the phase variation has the range of 2π, allowing for the estimation of frequency components related to vital signals. The respiration and heart rate in the baseband signals of the bio-radar can be obtained through IMF extraction using empirical mode decomposition, but the extraction level can vary depending on the surroundings and measurement conditions. The proposed method accurately extracts the IMFs at levels that contain the respiration and heart rate signals, and it helps in rejecting respiration harmonics located around the heart rate signal. Demonstrations using an in-house 5.8-GHz bio-radar sensor with the proposed method showed mean error rates of 4.2% and 2.1% in respiration and heart rate detection, respectively.