CAREER: Data Representation and Modeling for Unleashing the Potential of Multi-Modal Wearable Sensing Systems

职业：释放多模态可穿戴传感系统潜力的数据表示和建模

• 批准号: 1552828
• 负责人: Edgar Lobaton
• 金额: $ 49.21万
• 依托单位: North Carolina State University
• 依托单位国家: 美国
• 项目类别: Continuing Grant
• 财政年份: 2016
• 资助国家: 美国
• 起止时间: 2016-04-01 至 2022-03-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=1552828&HistoricalAwards=false
• 关键词: CAREER; Data; Representation; Modeling; Unleashing

The recent increase in the variety and usage of wearable sensing systems allows for the continuous monitoring of health and wellness of users. The output of these systems enable individuals to make changes to their personal routines in order to minimize exposures to pollutants and maintain healthy levels of exercise. Furthermore, medical practitioners are using these systems to monitor proper activity levels for rehabilitation purposes and to monitor threatening conditions such as heart arrhythmias. However, there is substantial work to be done to facilitate the processing and interpretation of such information in order to maximize impact. This proposal develops a computational framework that models the complex interactions between physiological and environmental factors contributing to an individual's health. The contributions of this award will facilitate the broad adoption of wearable sensing platforms and innovative analytical tools by individuals and medical practitioners.This award develops methodology for the estimation and prediction of physiological responses and environmental factors, with the objective of enabling users to efficiently change their behavior. To accomplish this objective, the framework will build on tools from statistical analysis, topological data analysis, optimization theory and human behavior analysis. This novel framework will not only develop new formal techniques, but it will also serve as a bridge between these cross-disciplinary fields. In particular, the proposed hierarchical computational framework has the potential of providing a trade-off between accuracy and computational flexibility based on the choice of granularity of the representation. This award will: (1) develop methodology for the concurrent representation of physiological, kinematic and environmental states for inference purposes; (2) develop techniques for mapping representations between different systems to enable information sharing; and (3) develop techniques to maximize the impact on the behavior of individuals by building on the proposed data representation. The algorithm development will be informed by integration of limitations on embedded platforms due to memory, computational and power capabilities, and transmission costs when off-board processing is required. The proposed techniques will empower users and medical practitioners to understand, analyze, and make decisions based on patterns in the data. The outcomes of this project will empower medical practitioners by providing innovative and effective tools for wearable sensing systems which enable efficient pattern identification, data representation and visualization. Besides training students directly working on this project, the data sets and algorithms developed will be incorporated into a new graduate course on computational techniques for physiological and environmental sensing. Undergraduate students will be engaged by participating in data collection experiments, REUs, and local demonstrations. Underrepresented undergraduate student communities will be exposed to the research at the national level by presenting demos at well-known diversity conferences in the STEM fields. Furthermore, K-12 local student communities will be engaged via summer workshops that will be prepared for students and educators.

最近可穿戴感测系统的种类和使用的增加允许持续监测用户的健康和健康。 这些系统的输出使个人能够改变他们的个人惯例，以尽量减少污染物的暴露，并保持健康的运动水平。此外，医疗从业者正在使用这些系统来监测用于康复目的的适当活动水平，并监测诸如心律失常的威胁状况。然而，要促进处理和解释这类信息，以最大限度地发挥影响，仍有大量工作要做。该建议开发了一个计算框架，该框架模拟了有助于个人健康的生理和环境因素之间的复杂相互作用。该奖项的贡献将促进个人和医疗从业者广泛采用可穿戴传感平台和创新分析工具。该奖项开发了用于估计和预测生理反应和环境因素的方法，旨在使用户能够有效地改变他们的行为。为了实现这一目标，该框架将建立在统计分析、拓扑数据分析、优化理论和人类行为分析等工具的基础上。这种新颖的框架不仅将开发新的形式化技术，而且还将作为这些跨学科领域之间的桥梁。特别是，所提出的分层计算框架具有提供基于表示的粒度的选择的准确性和计算灵活性之间的权衡的潜力。该奖项将：（1）开发用于同时表示生理、运动和环境状态的方法，以用于推理目的;（2）开发用于在不同系统之间映射表示的技术，以实现信息共享;以及（3）开发通过建立在所提议的数据表示的基础上来最大化对个人行为的影响的技术。 算法开发将通过集成嵌入式平台上的限制来通知，这些限制是由于内存、计算和功率能力以及需要板外处理时的传输成本。所提出的技术将使用户和医疗从业者能够根据数据中的模式来理解、分析和做出决策。该项目的成果将通过为可穿戴传感系统提供创新和有效的工具来增强医疗从业人员的能力，这些工具可以实现有效的模式识别，数据表示和可视化。除了培训直接从事该项目的学生外，开发的数据集和算法将被纳入一门关于生理和环境传感计算技术的新研究生课程。本科生将参与数据收集实验，雷乌斯，和当地的示范。代表性不足的本科生社区将通过在STEM领域的知名多样性会议上展示演示来接触国家层面的研究。此外，K-12当地学生社区将通过为学生和教育工作者准备的夏季研讨会参与。