CAREER: Automating the Construction and Analysis of Models of Motion for Prehabilitative Care

该学院早期职业发展计划(Career)项目将研究新的非侵入性方法，以自动识别前关键韧带(ACL)损伤的高风险个体。尽管训练有素的理疗师的指导护理已被证明能有效地预防此类伤害，但可用于此类再培训的资源有限。不幸的是，对这些韧带撕裂风险最高的患者的诊断依赖于训练有素的临床医生使用已被证明具有挑战性的定性指标进行昂贵、耗时的观察，这些指标已被证明难以一致应用。该项目将开发自动化的康复技术，通过构建和定量分析个人特有的肌肉骨骼运动模型来诊断那些前交叉韧带损伤风险增加的人。这将确保本项目中创建的诊断技术的可靠行为。重要的是，这种可广泛部署的方法将为前交叉韧带损伤的治疗增加一个预防性组成部分，这种损伤每年困扰着20多万人，并极大地影响患者此后的生活质量。更广泛地说，这些运动模型将从根本上改变人类辅助设备的控制方式。在控制过程中利用基于系统的技术的能力将使此类辅助设备能够广泛分布，目前这些辅助设备需要医生和工程师指导的仔细调整。该项目促进了科学的进步，为促进全民健康做出了重大贡献。综合教育计划不仅对本科教育有直接影响，而且通过亲身实践让K-12学生接触机器人控制，为他们提供了探索STEM职业生涯的极好机会。该项目支持开发数字技术，以诊断ACL损伤的风险增加，因为从一组摄像头观察到个人正在进行功能运动屏幕的关节姿势。为了构建和验证这些方法的可靠行为，本项目将探索以下三个研究目标：第一，一个新的凸优化工具，用于识别解释给定观测集的混合动力学模型的所有参数和相关控制输入。第二，一种新的诊断工具，可在所有已识别的模型参数中轻松评估伤害风险指标，以确保不会发生误诊。最后，对25名健康受试者和125名前交叉韧带重建手术后正在接受康复治疗的患者进行了为期两年的真实情况的纵向评估。这些在动力学、控制和优化方面的创新将使一种完全自动化、健壮、个性化和特定于运动的技术的开发成为可能，以识别那些面临ACL损伤风险增加的人，这是目前使用人工指导的方法无法实现的。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。

This Faculty Early Career Development Program (CAREER) project will investigate new non-invasive methods to automate the identification of individuals at elevated risk for Anterior Crucial Ligament (ACL) injuries. Though directed care by trained physical therapists has been shown to effectively prevent such injuries, the resources available for such re-training are limited. Unfortunately, the diagnosis of those at greatest risk for such ligament tears depends upon expensive, time-consuming observations by trained clinicians using qualitative metrics that have been shown to be challenging to consistently apply. This project will develop automated prehabilitative techniques to diagnose those at increased risk for ACL injury by constructing and quantitatively analyzing an individual-specific musculoskeletal model of motion. This will ensure the reliable behavior of the diagnostic technique created in this project. Importantly, this broadly deployable approach will add a preventive component for the treatment of ACL injuries that afflict more than 200,000 people annually and drastically affect a patient's quality of life thereafter. More broadly, these models of motion will fundamentally change how human assistive devices are controlled. The ability to leverage systems-based techniques during control will enable the wide distribution of such assistive devices, which currently require careful physician and engineer-guided tuning. This project promotes the progress of science and contributes significantly to advance the national health. The integrated education plan not only has direct impact on undergraduate education but also provides great opportunities for K-12 students to explore STEM careers by exposing them to robotic control via hands-on examples.This project supports the development of numerical techniques to diagnose an increased risk for ACL injury given articulated pose observations from a set of cameras of an individual performing a functional motion screen. To construct and verify the reliable behavior of these methods, this project will explore the following three research objectives: First, a new convex optimization tool to identify all parameterizations and associated control inputs of a hybrid dynamical model that explain a given set of observations. Second, a novel diagnostic tool that tractably evaluates injury risk metrics across all identified model parameterizations to ensure that misdiagnosis does not occur. Finally, a real-world, two yearlong longitudinal evaluation on 25 healthy subjects and 125 patients who are undergoing rehabilitation after an ACL reconstructive surgery. These innovations in dynamics, controls, and optimization will enable the development of a completely automated, robust, individual, and motion specific technique to identify those at increased risk for ACL injuries that cannot be currently achieved with the existing human-guided approach.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.