Yang Chenxi, Bruce Brendan, Liu Xiaofan, Gholami Behnood, Tavassolian Negar
Annu Int Conf IEEE Eng Med Biol Soc. 2020 Jul;2020:2646-2649. doi: 10.1109/EMBC44109.2020.9175267.
This paper reports a pilot study of a hybrid radar-camera system that simultaneously monitors the respiration of two subjects. A prototype system was built involving a low-cost impulse-radio ultra-wideband (IR-UWB) radar module and an optical and depth-sensing camera module. The system detects subjects using the camera and utilizes the distance information acquired to guide the signal processing of the radar. This structure simplifies subject identification and tracking for the radar system, provides further context to the radar, and facilitates the extraction of respiration information. Experiments under different scenarios were conducted to evaluate the performance of the system at different distances and angles from subjects. The localization procedure has an average accuracy of 0.1 m. The respiration rates extracted from the radar are comparable with the values from the reference wearable sensor, reporting an average error of 0.79 respirations per minute (RPM) with a standard deviation of 0.71 RPM. The results suggest that the respiration signals from subjects could be accurately monitored with the presented framework. It is also feasible to monitor two subjects at the same time in most scenarios. The proposed framework shows promising potential to work as a ubiquitous monitoring system for multiple subjects.
本文报道了一项关于混合雷达-相机系统的初步研究，该系统可同时监测两名受试者的呼吸。构建了一个原型系统，其中包括一个低成本的脉冲无线电超宽带（IR-UWB）雷达模块以及一个光学和深度感应相机模块。该系统使用相机检测受试者，并利用获取的距离信息来指导雷达的信号处理。这种结构简化了雷达系统的受试者识别和跟踪，为雷达提供了更多背景信息，并有助于呼吸信息的提取。在不同场景下进行了实验，以评估系统在与受试者不同距离和角度时的性能。定位程序的平均精度为0.1米。从雷达提取的呼吸率与参考可穿戴传感器的值相当，平均误差为每分钟0.79次呼吸（RPM），标准差为0.71 RPM。结果表明，利用所提出的框架可以准确监测受试者的呼吸信号。在大多数情况下同时监测两名受试者也是可行的。所提出的框架显示出作为多受试者通用监测系统的巨大潜力。
