SCH: EXP: RadiOptiMeter: Long-Term and Fine-Grained Breathing Volume Monitoring for Sleep Disordered Breathing (SDB)

儿童睡眠呼吸紊乱(SDB)被认为是一种公共健康问题，其严重后果包括认知功能下降、学习成绩差、白天嗜睡和心血管疾病风险增加。目前SDB的诊断是在医院睡眠实验室通过在患者S身体的不同位置安装大量心肺传感器来监测患者的。这种突出的监测形式给患者带来了不便，需要技术人员投入大量精力来确保研究质量。尤其是儿童，他们对这项研究的耐受性很差；经常移除传感器，睡眠困难，需要重复检查。该项目旨在开发一种新的方法，在睡眠研究期间使用光学信号辅助射频信号来远程连续监测受试者的呼吸量和呼吸模式。我们介绍了RadiOptiMeter，一种混合的射频-光学呼吸量监测方法，将射频(RF)信号的独特特征与深度-CO2-热像仪捕获的图像流结合在一起，以准确估计远处睡眠患者的呼吸量。这项研究是发展无创呼吸监测的第一步。利用这种新颖的非侵入性设备测量睡眠中的呼吸将启动一项研究计划，以促进未来在儿童舒适的S家中诊断SDB，而不需要住院实验室费用或每个睡眠实验室每年相当于数千美元的一次性医疗设备的费用。该项目将探索一种新的方法，在医院睡眠研究期间使用射频和光学信号连续监测人类的呼吸量和呼吸模式。我们引入了RadiOptiMeter，这是一种无线-光学混合呼吸量监测方法，将射频(RF)信号的独特特征与深度-CO2-热像仪捕获的图像流结合在一起，以准确且连续地估计远处睡眠患者的呼吸量。我们提出了解决睡眠期间身体运动、环境无线信号噪声和患者群体多样性带来的挑战的技术。一个预期的结果是一个强大和准确的呼吸量监测系统，用于SDB研究。建议的每个设备之间的合作使我们能够利用设备的协同效应来覆盖每种设备类型施加的限制，并提供系统冗余。这些冗余确保了长期监测任务的可靠性，而长期监测任务对临床应用至关重要。我们的研究将为实现新的非接触式生命信号监测系统做出以下关键贡献：(1)基于视觉系统(VVE)的呼吸量估计方法的分析模型、实验工具和评估结果，其中包括从深度-CO2-热像仪(DCT)输出的4D体积模型和骨骼结构分析。(2)基于射频的呼吸量估计(RVE)系统的分析模型、实验硬件和软件组成以及评价结果，该系统使用基于神经网络的机器学习进行胸部位移-体积匹配。(3)将VVE和RVE协同结合的射光混合呼吸量估计系统(RadiOptiMeter)，以进行连续和细粒度的监测。RadiOptiMeter包括身体运动跟踪、自动天线转向和一套控制和同步算法，实现了整个系统的和谐集成。科罗拉多州S儿童医院计算机科学与工程系的研究人员和睡眠医学研究的医生们的合作是非侵入性呼吸监测发展的第一步。一种新型的非侵入性睡眠呼吸测量设备将是必要的基础研究的第一步，以促进未来在儿童舒适的S家中诊断SDB，而不需要住院实验室费用或每个睡眠实验室每年相当于数千美元的一次性医疗设备的费用。该项目也为培养研究生进行基于视觉的研究项目提供了一种很好的方法。RadiOptiMeter的概念是一种令人兴奋和有吸引力的工具，可用于组织各种教育活动。此外，项目成果将通过学术出版物传播，并通过我们现有和潜在的工业合作伙伴进行积极推广。

Sleep disordered breathing (SDB) in children is considered to be a public health problem with serious consequences such as decreased cognitive function, poor school performance, daytime sleepiness and increased cardiovascular risk. Current diagnosis of SDB is performed in hospital sleep laboratories by monitoring patients with a host of cardiorespiratory sensors attached at various positions on the patient?s body. This obtrusive form of monitoring is inconvenient to patients and require tremendous amount of attention from technicians to ensure study quality. Children, especially, tolerate the study very poorly; often removing sensors, having trouble sleeping, and necessitating repeat investigations. This project aims to develop a new method to remotely and continuously monitor breathing volume and breathing patterns of human subjects during sleep studies using optical signals assisted by radio frequency signals. We introduce RadiOptiMeter, a hybrid radio-optical breath volume monitoring approach that couple the unique characteristics of radio frequency (RF) signals with image stream captured by a depth-CO2-thermal camera to accurately estimate breathing volume of sleeping patients from afar. This study is the first step in the development of non-invasive respiratory monitoring. Utilizing this novel and non-invasive device to measure breathing during sleep will begin a research program to promote future diagnosis of SDB in the comfort of the child?s home, with no in-hospital laboratory expenses or the expense of the disposable medical equipment which equates to thousands of dollars a year in each sleep laboratory. This project will investigate a new method to continuously monitor breathing volume and breathing patterns of humans during in-hospital sleep studies using radio frequency and optical signals. We introduce RadiOptiMeter, a hybrid radio-optical breath volume monitoring approach that couple the unique characteristics of radio frequency (RF) signals with image stream captured by a depth-CO2-thermal camera to accurately and continuously estimate breathing volume of sleeping patients from afar. We propose techniques to address challenges brought about by the body movement during sleep, environmental wireless signal noises, and the diversity of patient populations. An expected outcome is a robust and accurate breathing volume monitoring system for SDB studies. The cooperation between each of the proposed devices allows us to exploit the synergistic effects of the devices to cover the limitations imposed by each device type and provides system redundancies. These redundancies ensure reliability for long-term monitoring tasks, which are critical for clinical applications. Our proposed research will make the following key contributions to enable new non-contact vital signal monitoring system: (1) Analytical models, experimental tools, and evaluation results of a breathing volume estimation method using a vision-based system (VVE) which include 4D volumetric model and skeletal structure analysis from depth- CO2-thermal (DCT) camera outputs. (2) Analytical models, experimental hardware and software components, and evaluation results of a RF-based breathing volume estimation (RVE) system, that uses neural-network-based machine learning for chest displacement- to-volume matching. (3) A hybrid radio-optical breathing volume estimation system (RadiOptiMeter) that synergistically combines the VVE and RVE to perform continuous and fine-grain monitoring. RadiOptiMeter includes body movement tracking, automatic antenna steering, and a set of controlling and synchronizing algorithms for a harmonic integration of the whole system. This collaborative effort between researchers at the Department of Computer Science and Engineering and medical doctors at Sleep Medicine Research at the Children?s Hospital Colorado is the first step in the development of non-invasive respiratory monitoring. A novel non-invasive device to measure breathing during sleep will be the first step in the necessary foundational research to promote future diagnosis of SDB in the comfort of the child?s home, with no in-hospital laboratory expenses or the expense of the disposable medical equipment which equates to thousands of dollars a year in each sleep laboratory. This project also provides an excellent methodd to train graduate students to conduct this vision-based research project. The RadiOptiMeter concepts serves as an exciting and appealing tool for structuring a variety of educational activities. Moreover, the project results will be disseminated through scholarly publications and active outreach through our existing and potential industrial partners.