Human Identification by Measuring Respiration Patterns Using Vital FMCW Radar

This letter proposes a method of human identification that measures respiration patterns using frequency-modulated continuous wave (FMCW) radar. We exploit the fact that respiration signal patterns are unique to each individual, and FMCW radar is employed to obtain the respiration information. Based on the strengths of FMCW radar, the proposed algorithm compensates for the inability to distinguish the respiration signals of multiple users, which are difficult for continuous wave radar to measure. The proposed algorithm also employs a deep neural network algorithm instead of the K-nearest neighbor algorithm that was used in a previous study. The proposed algorithm further improves the performance by using a least mean square filter in the input signal of the DNN. The experimental results show that the proposed human identification method successfully classified four persons.