Collaborative Research: FW-HTF-R: Future of Construction Workplace Health Monitoring

鉴于建筑业的伤亡比例过高，以及迫在眉睫的技术革命和气候变化可能带来不明确的健康和危险情况，至关重要的是改善未来劳动力的健康和安全。然而，缺乏一种有效的、客观的和持续的方法来评估建筑工人在工地的健康状况。虽然可穿戴式生理传感技术和人工智能在客观评估建筑工人的健康参数方面已经有了重要的创新，但建立以工人为中心、具有良好预防潜力的整体健康监测方法仍然存在根本挑战。这些挑战来自：a)缺乏可扩展和可行的可穿戴传感器来持续激发工人对现场应激源的不同身体反应；b)缺乏强大的解释性数据驱动框架来处理所产生的信号，以便自动早期检测身体疲劳、精神压力和暴露在热应激下；以及c)缺乏有效的向工人和经理提供健康和安全信息的表示，以便通过增强他们的情景意识来改进任务决策。通过建立一种实时和背景感知的整体健康监测方法，该项目将在改善美国建筑业近700万工人的安全方面发挥基础性作用。预计开发的智能健康监测系统将改变工作质量和劳动力政策，从而减少冲突，提高生活质量。它还可用于解决其他危险行业的工作场所健康问题，如制造业、消防和农业。本研究的总体目标是通过整合灵活、可穿戴传感器制造、人工智能和隐私感知信息可视化方面的多学科研究，为工人和管理人员提供近乎实时和预测的未来情景感知健康和安全信息，从而通过增强他们的情景意识来改进任务决策。这个项目的智力意义在于通过在三个方面产生和扩展新知识来实现目标。首先，该项目将设计和制造一种柔性可穿戴传感器，用于连续和非侵入性地测量工人在建筑工地上的生物电信号和电化学反应。使用单一、灵活的可穿戴传感设备而不是多个现成的传感器，将有助于拟议的健康传感系统在建筑工作场所的可扩展性和可行性。其次，该项目将开发强大的机器学习算法和框架，以基于生理、背景和环境数据对现场工人的健康状况进行持续和客观的评估。为此，该项目将通过开发一种新的解释性数据驱动方法来解决与传统机器学习算法有关的基本挑战，该方法在确保数据安全和隐私的同时，对个人间和个人内部的可变性具有很强的抵抗力。第三，这项研究将通过一系列集体健康分析生成建筑工地的数字孪生模型(健康和安全地图)，并开发一个自动反馈模块，向外地工作人员提供与个人健康有关的信息和相应的缓解战略。对集体健康和安全信息的洞察可以深刻地帮助工人和安全经理近乎实时地做出关于执行以现场为导向的建筑作业的合理、有远见的决策。这项研究工作将通过集体可视化工人的实时健康和安全信息，为改进现场的主动健康和安全管理打开新的大门。该奖项反映了NSF的法定使命，并通过使用基金会的智力优势和更广泛的影响审查标准进行评估，被认为值得支持。

Given the disproportionate rate of fatalities and injuries in the construction industry and the potential of ambiguous health and hazardous situations with respect to the impending technological revolution and climate change, it is crucial to improve the health and safety of the future workforce. However, there is a lack of an effective, objective, and continuous approach for assessing construction workers' health status at jobsites. Although there have been important innovations in wearable physiological sensing technologies and artificial intelligence for objective assessment of construction workers' health parameters, there remain fundamental challenges for establishing a worker-centered holistic health monitoring approach with promising preventive potentials. These challenges stem from: a) lack of a scalable and feasible wearable sensor for continuous elicitation of workers' diverse bodily responses to stressors in the field; b) lack of a robust interpretive data-driven framework to process the elicited signals for automatic early detection of physical fatigue, mental stress, and exposure to heat stress; and c) lack of effective representation of health and safety information to workers and managers for enabling improved task decisions by augmenting their situational awareness. By establishing a real-time and context-aware holistic health monitoring approach, this project will play a fundamental role in improving the safety of close to 7 million workers in the U.S. construction sector. The developed intelligent health monitoring system is expected to produce changes in the quality of work and workforce policies, resulting in reduced conflicts and enhanced quality of life. It can also be used to address workplace health issues in other hazardous industries such as manufacturing, firefighting, and agriculture.The overarching goal of this research is to improve construction workforce health and safety by integrating multi-disciplinary research in flexible, wearable sensor fabrication, artificial intelligence, and privacy-aware information visualization to provide near-real-time and projected future context-aware health and safety information to workers and managers for enabling improved task decisions by augmenting their situational awareness. The intellectual significance of this project lies in fulfilling the goal by generating and expanding new knowledge on three fronts. First, the project will design and fabricate a flexible wearable sensor for continuous and noninvasive measurement of workers' bioelectric signals and electrochemical responses at construction sites. The use of a single, flexible wearable sensing device instead of multiple off-the-shelf sensors will facilitate the scalability and feasibility of the proposed health sensing system in the construction workplace. Second, the project will develop robust machine learning algorithms and frameworks for continuous and objective assessment of workers' health conditions in the field based on physiological, contextual, and environmental data. For this purpose, this project will address fundamental challenges related to traditional machine learning algorithms by developing a novel interpretive data-driven approach robust to inter- and intra-individual variability while ensuring data security and privacy. Third, this research will generate a digital twin model (health and safety maps) of the construction sites through an array of collective health analyses and develop an automated feedback module for providing personal health-related information and corresponding mitigation strategies to field workers. The insights into the collective health and safety information can profoundly assist the workers and safety managers in making a sound, far-sighted decision about the execution of field-oriented construction operations in near real-time. This research effort will open new doors in improving proactive health and safety management in the field through collective visualization of workers' real-time health and safety information.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.