EAGER: Preliminary Studies for Real-Time Tracking of Cognitive Fatigue via Eye Movement Patterns

本项目旨在对端到端mHealth系统的设计进行初步探索，以利用眼球运动来检测个体的疲劳状态。疲劳状态的实时检测对于减少医疗差错、减少机动车操作员差错以及改善公众的整体健康和健康状况等广泛应用非常重要。尽管有这些好处，但还没有广泛可用的方法来预测自然环境中的疲劳。在这个项目中，我们正在设计一个实时分析引擎，通过融合关于各种眼睛测量(扫视、注视、眨眼和瞳孔放大)的原始时间序列数据，提供推断疲劳的通用计算方法。这项工作包括一种针对自然环境的新型疲劳检测器，该检测器解决了对真实世界变异性、个体差异和混杂因素的稳健性，并开发了时间序列模式提取方法来表征疲劳状态的时间进程。我们还在为学生群体对我们的设计进行试点评估。我们预计，一个成功的试点将导致更广泛的研究，涉及住院医生、机动车辆操作员、轮班工人和其他疲劳是一个重要的安全或健康问题的用户群体。该项目包括为中学生举办的为期两天的移动传感主题研讨会。研讨会在马萨诸塞州大学阿默斯特分校举办，建立在与马萨诸塞州霍利奥克的Girls Inc.现有的外展伙伴关系的基础上。这是为了解决在科技、经济和经济领域普遍认识到的妇女任职人数不足的问题。

This project seeks to perform a preliminary exploration into the design of an end-to-end mHealth system for detecting the fatigue state among individuals by leveraging eye movements. Real-time detection of fatigue state is important for a wide range of applications including reducing medical errors, reducing motor vehicle operator errors, and improving overall health and wellness for the general public. Despite these benefits, there is no widely available method for predicting fatigue in the natural environment.In this project, we are designing a real-time analytics engine that provides general computational methods for inferring fatigue by fusing raw time-series data about various eye measures (saccades, fixations, blinks, and pupil dilation). This work includes a novel fatigue detector for natural settings that address robustness to real-world variability, individual differences, and confounders, and develops time-series pattern extraction methods to characterize temporal progression of fatigue state. We are also performing a pilot evaluation of our design for a student population. We envisage that a successful pilot will lead to more extensive studies involving medical residents, motor vehicle operators, shift workers, and other user groups where fatigue is a significant safety or health issue. This project includes outreach including a two day mobile sensing-themed workshop for middle school students. The workshop is offered at the UMass Amherst campus and builds on an existing outreach partnership with Girls Inc. of Holyoke, MA. that is aimed at addressing the widely recognized under-representation of women in STEM fields.