In-Car Multiple Targets Vital Sign Monitoring Using Location-Based VMD Algorithm

Abstract

In-car vital sign monitoring is of great importance for ensuring safe driving. Impulse Radio Ultra-WideBand (IRUWB) radar with low power consumption and high resolution, is widely used for vital sign monitoring. However, existing researches focus on monitoring the single target in the car, especially for the driver. This paper proposes a location-based Variational Mode Decomposition (VMD) algorithm for in-car vital sign monitoring on multiple targets, including the driver and passengers, with only one radar. Firstly, the prior knowledge of seat locations is obtained after determining the position of the radar, based on which the received radar signals are divided into several intervals. Secondly, VMD is performed on each signal interval to extract vital signals in the corresponding location. Finally, in order to alleviate the severely superposed signals caused by the multipath environment in the car, the correlation coefficients between extracted vital signals in different locations are compared to select the corresponding vital signs. The Heart Rate (HR) and Respiration Rate (RR) of the driver and passengers are obtained. In addition, this paper analyzes the impact of passengers' locations when monitoring multiple targets vital signs. To the best of our knowledge, this is the first attempt to monitor multiple targets in the car. Experiments are conducted in an off-the-shelf car with five seats, and two scenarios, including only the driver and two targets (the driver and the passenger), are considered. Then real time monitoring is implemented. Experimental results demonstrate the effectiveness of the proposed algorithm. The accuracies of vital sign monitoring on only the driver and two targets reach 92.66% and 88.44% respectively.