SenSE:Wearable hybrid biochemical and biophysical sensing systems integrated with robust artificial intelligence for monitoring COVID-19 patients

截至二零二零年七月，二零一九年冠状病毒病（COVID-19）爆发已感染全球逾1，700万人，导致逾669，000人死亡。根据现有数据和已发表的报告，大多数被诊断患有COVID-19的人没有或表现出轻微症状，可以出院回家进行自我隔离。其中约20%会发展为需要住院和医疗管理的严重疾病。目前，医疗保健提供者缺乏有效的方法和技术来在家中远程监测患者的临床状况，评估他们的疾病进展，并预测临床恶化以进行及时的医疗干预。该多学科项目旨在通过提供家庭智能监控系统，为改善COVID-19复苏成果创造新途径。该项目将提供令人兴奋的跨学科教育和研究机会，以及实践经验，以培养我们的研究生，并让本科生，特别是少数民族学生参与研究。本集团将推行一系列综合研究及教育活动，以外展至K-12学生及公众，提高他们对先进科学及工程解决方案的认识，以应对COVID-19所带来的重大医疗挑战。最近的研究已经确定，细胞因子水平与COVID-19疾病的严重程度和死亡率相关。细胞因子水平可作为疾病严重程度和进展的有效预测因子。目前，细胞因子的测试涉及有组织的设置，例如由采血师进行血液采集，并使用实验室设备（例如读板器）分析样品。一个主要的缺点是，不能实现细胞因子水平的连续监测，因为它需要随着时间的推移到医院进行数十次访问。该提案的目标是开发一种可穿戴的多模式传感系统，集成可解释和强大的人工智能，用于持续监测COVID-19患者的生物物理和生化状况，密切跟踪其病情进展，并及时预测风险水平和医疗干预。该项目包括三个研究目标：（1）开发可穿戴生物化学/生物物理传感系统，用于COVID-19患者的非侵入性和连续监测，（2）将可穿戴传感系统与可解释和强大的人工智能集成，用于多模式传感器数据分析，个性化疾病进展建模和传感器性能优化，以及（3）表征和评估具有COVID-19患者的多模态传感器系统。这项研究将为开发一种用于长期和连续监测细胞因子的新型传感器提供基本的理解和基本原则。将开发先进的机器学习方法和工具，用于多模态传感器数据分析、风险水平确定和传感器性能优化。该项目开发的生物传感技术、设备设计和机器学习模型适用于其他领域，包括监测流感或其他疾病患者的传感器，这些领域需要持续监测和及时干预。该奖项反映了NSF的法定使命，并通过使用基金会的智力价值和更广泛的影响审查标准进行评估，被认为值得支持。

The outbreak of Coronavirus Disease 2019 (COVID-19) has infected more than 17 million individuals worldwide, resulting in the death of more than 669, 000 people as of July 2020. Based on the available data and published reports, most people diagnosed with COVID-19 exhibit no or mild symptoms and could be discharged home for self-isolation. About 20% of them will progress to severe disease requiring hospitalization and medical management. Currently, there is a lack of effective methods and technologies for healthcare providers to remotely monitor patients’ clinical conditions at home, evaluate their disease progression, and predict clinical deterioration for timely medical interventions. This multidisciplinary project aims to create a new route to improve the COVID-19 recovery outcome by providing an at-home smart monitoring system. This project will provide exciting interdisciplinary education and research opportunities, as well as hands-on experience, to train our graduate students and to involve undergraduate students, especially minority students, into research. A set of integrated research and education activities will be implemented for out-reaching to K-12 students and the public, to increase their awareness of advanced scientific and engineering solutions for addressing the critical healthcare challenges of COVID-19. Recent studies have established that cytokine level is associated with COVID-19 disease severity and mortality. The level of cytokines can be used as an effective predictor for disease severity and progression. Currently, testing for cytokines involves an organized setup such as blood collection by a phlebotomist and analysis of samples using laboratory equipment such as plate readers. A major drawback is that continuous monitoring of cytokine levels cannot be accomplished since it will require dozens of visits to the hospital over time. The objective of this proposal is to develop a wearable multimodal sensing system integrated with explainable and robust artificial intelligence for continuous monitoring of biophysical and biochemical conditions of COVID-19 patients at home, close tracking of their illness progression, and timely risk level prediction and medical intervention. This project includes three research objectives: (1) develop a wearable biochemical/biophysical sensing system for non-invasive and continuous monitoring of COVID-19 patients, (2) integrate wearable sensing system with explainable and robust artificial intelligence for multimodal sensor data analysis, personalized illness progression modeling, and sensor performance optimization, and (3) characterize and evaluate the multimodal sensor systems with COVID-19 patients. The research will provide fundamental understanding and essential principles for developing a novel sensor for long-term and continuous monitoring of cytokines. Advanced machine learning methods and tools will be developed for multimodal sensor data analysis, risk level determination, and sensor performance optimization. The biosensing technology, device design, and machine learning models developed in this project are applicable to other fields, including sensors to monitor patients with influenza or other diseases, where the continuous monitoring and timely interventions are required.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.