Heart Rate Variability Monitoring under Stimulation Input Using Non-Contact CW Radar

Abstract

Recent studies have mainly relied on radar technology for extracting crucial vital signs because of its ability to measure without physical contact. The focus on deriving cardiac interbeat interval and heart rate variability has gained significance due to its complexity and relevance in healthcare. Our investigation involved a detailed analysis of continuous wave radar signals to enhance the extraction of chest wall movement data. Using a convolution algorithm, we eliminated the respiratory component from the signal, while a locally projective noise reduction algorithm helped isolate the heartbeat component. Subsequently, a derivation filter was applied to pinpoint the R peak of the heartbeat, facilitating the collection of IBI and HRV metrics. This methodology proved effective for individuals in a relaxed, motionless state. However, its efficacy in cases of elevated heart rates caused by factors such as exercise or caffeine consumption remained uncertain. For subjects with large changes in heart rate followed by large changes in cardiac IBI, we made a small improvement in the algorithm. By performing a window shift of 10 seconds with an overlap of 1 second. Each 10-second data segment is fed into the algorithm. At each data segment, we perform multiple iterations with decreasing number of neighbors until no further change is made. With this adjustment, the results achieved in the group of subjects using input stimulation to increase heart rate such as exercise, or drinking coffee were indicated a strong correlation of 96.38% between radar-based measurements and reference measurements for this group of subjects, affirming the effectiveness of the proposed method in such scenarios.