HCC: SMALL: Wearable computation and feedback for real-time movement training

运动是人类生存的基本特征，与我们的生活质量密切相关。 运动训练可以预防损伤，提高运动成绩，延缓肌肉骨骼疾病，加速康复。 到目前为止，培训的范围仅限于专门的培训设施，效果也仅限于体能教练的口头建议。 在这个项目中，PI的目标是扩大人体运动训练的范围和有效性，以扩大健康和生活方式的好处，以广大公众。 这项研究在机器人、生物力学和人机交互交叉领域的主要成果将是步态建模软件，该软件将真实的传感器数据集成在一起，计算运动学、动力学和关节/肌腱力，以预测对运动参数的调整，从而实现最终目标。 PI的假设是，可穿戴计算可以从根本上改变人们的移动方式。 计算硬件的小型化，以及运动分析算法、可穿戴传感器和反馈设备的进步，都是人类互动新水平的催化剂。 这项工作将侧重于日常重复的动态活动，如步行，跑步和跳跃，其本质是在一个周期中收集和分析的信息可以应用于后续周期，以实现逐步改善。 将以真实的时间提供建立在机器人技术、运动跟踪和生物力学建模（已导致对复杂的多自由度系统的有效监测和模拟）的进步基础上的反馈，并且该反馈将是自适应的、相对于周期到周期的变化是鲁棒的并且是用户特定的。 为了达到最大的效果，运动训练将面向普通公民，而不是局限于实验室或诊所。 为此，PI将创建一个系统，可以在房子周围走动，户外徒步旅行或在健身房跑步时使用。 在运动跟踪，动态分析和可穿戴反馈步态再训练，以减少与损伤和关节炎相关的膝关节负荷的初步实验将扩展到评估哪种类型的传感和反馈，结合算法来检测和分析运动异常，是有效的实验室外。 PI将进行一系列实验来验证便携式解决方案，将其与在完全仪器化的实验室环境中获得的结果进行比较，并评估培训效果如何随着时间的推移而保留。 更广泛的影响：PI认为，通过可穿戴再训练设备，中年妇女可以被教导以一种减缓或预防骨关节炎以及臀部、脚踝等部位受伤的方式走路，大学生运动员可以在打排球时接受跳跃和落地训练，以防止前交叉韧带和其他常见的运动相关损伤（同时也可能提高成绩），中风和其他神经系统疾病的受害者可以在家中康复，而不是在诊所。 该项目将首先关注步行和膝关节负荷，这是美国老龄化人口的一个直接重要问题。 PI将提供开源软件（例如，用于监测传感器和预测目标步态参数）和可穿戴硬件许可，以促进项目成果适应其他应用。

Movement is a basic feature of human existence and is intimately connected to our quality of life. Movement training can prevent injury, improve athletic performance, delay musculoskeletal disease and accelerate rehabilitation. Until now, training has been limited in scope to specialized training facilities and limited in effectiveness to the verbal recommendations of physical trainers. In this project, the PI's goal is to expand the scope and effectiveness of human movement training in order to extend health and lifestyle benefits to the general public. The primary outcome of this research at the intersection of robotics, biomechanics and human-computer interaction will be gait modeling software that integrates sensor data in real time to compute kinematics, kinetics and joint/tendon forces to predict adjustments to motion parameters to achieve an end goal. The PI's hypothesis is that wearable computation can fundamentally change the way people move. The miniaturization of computational hardware, as well as advances in movement analysis algorithms, wearable sensors and feedback devices, all serve as catalysts for a new level of human interaction. The work will focus on everyday repetitive dynamic activities like walking, running and jumping, whose nature is that information gathered and analyzed during one cycle can be applied to subsequent cycles to achieve gradual improvement. Feedback that builds upon advances in robotics, motion tracking and biomechanical modeling (that have led to efficient monitoring and simulation of complex multi-degree of freedom systems) will be provided in real time and will be adaptive, robust with respect to cycle-to-cycle variations, and user specific. For maximum impact, movement training will be accessible to the average citizen instead of confined to the laboratory or clinic. To this end, the PI will create a system that could be used while walking around the house, hiking outdoors or running in a gymnasium. Preliminary experiments in motion tracking, dynamic analysis and wearable feedback for gait retraining to reduce knee loading associated with injury and arthritis will be extended to evaluate which types of sensing and feedback, in combination with algorithms to detect and analyze motion anomalies, are effective outside of the laboratory. The PI will conduct a series of experiments to validate the portable solution, comparing it with results obtained in a fully instrumented laboratory setting and assessing how the effects of training are retained over time. Broader Impacts: The PI argues that with wearable retraining devices middle-aged women could be taught to walk in a way that slows or prevents osteoarthritis as well as injury at the hips, ankles, etc., college athletes could be trained to jump and land while playing volleyball so as to prevent ACL and other common sports-related injuries (while perhaps improving performance as well), and victims of stroke and other neurological disorders could be rehabilitated at home instead of at the clinic. This project will focus initially on walking and knee joint loading, a problem of immediate importance for the aging U.S. population. The PI will provide open source software (e.g., for monitoring sensors and predicting target gait parameters) and wearable hardware licensing to promote adaptation of project outcomes to other applications.