Wearable technology to inform fatigue and injury processes with ergonomic applications

本文概述的研究将推进人体工程学、计算建模、上肢生物力学、反复劳损病因学和可穿戴技术等领域的发展。包括腕部运动学和指尖力在内的生物力学数据将在工作场所的长时间范围内(整个班次)收集。这些大型数据集将使用机器学习和分类技术进行检查，以寻找普遍趋势(即，检测导致损伤的疲劳)。这些数据还将为反向动力学建模提供信息，随后是模拟体外测试、高分辨率计算机断层成像和计算建模。该研究的结果有望为损伤风险的生物力学评估提供信息。此外，拟议的研究计划将产生一个公共存储库，其中包括上肢(手腕)运动学的大型数据集，以及使用定制的CT兼容关节模拟器通过相关运动模拟的身体手腕的高分辨率、三维、动态、计算机断层扫描图像。这些造影剂将用于身体手腕，以可视化骨骼和软组织结构。通过逆动力学计算计算的肌肉负荷将为模拟工作场所的运动和加载场景提供信息。然后，生成的图像将被用于开发预测组织应力和应变的计算模型。这些模型将通过使用数字图像相关方法对现场软组织应变的实验测量进行验证。整个研究计划将使用来自可穿戴设备的数据，实现对伤害风险的知情和验证预测，这些设备可以很容易地整合到工作场所。此外，这项研究将填补几个空白，包括将可穿戴数据与经过验证的生物力学分析联系起来；以及为常用的建模技术提供可靠的验证数据。最后，机器学习算法的应用有望在疲劳、伤害预防和生物力学风险计算方面产生新的发现。

The research outlined herein will advance the areas of ergonomics, computational modelling, upper extremity biomechanics, aetiology of repetitive strain injuries, and wearable technology. Biomechanical data including wrist kinematics and fingertip forces will be collected over long time frames (entire shifts) in a workplace study.  These large datasets will be examined using machine learning and classification techniques to look for generalized trends (ie., detection of fatigue leading to injury).  This data will also inform inverse dynamics modelling, followed by simulated in-vitro testing, high resolution computed tomography images and computational modelling.  The results of this research are expected to inform biomechanical evaluation of injury risk. In addition, the proposed research plan will produce a public repository of large data sets of upper extremity (wrist) kinematics, as well as high resolution, three dimensional, dynamic, computed tomography images of cadaveric wrists simulated through relevant motions using a custom CT compatible joint simulator.  Contrast agents will be used in the cadaveric wrists to visualize both bony and soft tissue structures.  Muscle loads, calculated through inverse dynamics calculations, will inform the simulation of workplace movements and loading scenarios. The resulting images will then be used to develop computational models that will predict tissue stresses and strains.  These models will be validated by experimental measure of in-situ soft tissue strain using digital image correlation methods.  The entire research plan will enable informed and validated prediction of injury risk, using data from wearable devices that can easily be incorporated in the workplace.  In addition, the research will serve to close several gaps, including linking wearable data to validated biomechanical analysis; and providing sound validation data to commonly used modelling techniques.  Finally, the application of machine learning algorithms is anticipated to yield new discovery around fatigue, injury prevention, and biomechanical risk calculations.