Sparse Processing for Driver Respiration Monitoring using In-Vehicle mmWave Radar

We investigate non-contact driver's breath-rate monitoring using a compact millimeter-wave (mmWave) in-vehicle radar. The small form-factor mmWave sensor allows deploying multiple antennas in space-constrained automotive applications while also improving the spatial resolution of the radar. We formulate the breathing rate extraction as a sparse and low-rank recovery problem. Our sparse spectral estimation approach then accurately extracts driver respiration rate from highly cluttered and noisy radar measurements. We validate our proposed approach through real-data collected from a 77 GHz TI AWR6843 radar sensor deployed in a driving simulator.