A Theoretical Investigation of the Detection of Vital Signs in Presence of Car Vibrations and RADAR-Based Passenger Classification

The observation of the passenger compartment using an interior RADAR becomes challenging when a vehicle is in motion. The signal received by the RADAR sensor measures not only the target's motion that we would like to detect, e.g., vital signs, but also a second, unwanted motion induced by exterior effects, e.g., displacement caused by the road, strong wind gusts, or car engine vibrations. Consequently, traditional techniques used to remove white Gaussian noise cannot be applied in this scenario. We propose a theoretical framework to investigate the denoising of a signal received by an interior RADAR using accelerometer measurements and a mechanical model to generate the cancellation signal stemming from the (unwanted) vibrational motion. We explain the denoising technique and show that the fundamental idea behind the cleaning process works well. Our observations are corroborated by a robust mathematical model and verified by simulations. Furthermore, the aforementioned denoising techniques enable us to formulate and solve an optimization problem to find an estimation for the passenger's mass in certain cases.