Random Body Movement Interference Mitigation in Radar Breath Detection Based on L1 Norm

Radar-based noncontact detection of vital signs holds significant advanta-ges over conventional measurement devices. However, random body motion (RBM) leads to a decrease in the accuracy of radar-based human respiratory frequency estimation. In order to tackle this challenge, in this letter, we propose an RBM interference mitigation framework based on the L1 norm, which can be delineated into three specific steps. In the first step, we normalize the amplitude of the preprocessed signal to detect the signal region affected by RBM interference. In the second step, we introduce a labeling technique to eliminate the regions affected by RBM interference. In the third step, we reconstruct the missing respiratory signal data using the L1 norm-based least squares method. The experiments were conducted in a real-world environment, and the results demonstrate that the proposed RBM interference mitigation method based on the L1 norm effectively reduces the impact of RBM on respiratory rate accuracy. In comparison to existing algorithms, it exhibits superior root-mean-square error and real-time monitoring performance.