CPS: Synergy: Collaborative Research: Closed-loop Hybrid Exoskeleton utilizing Wearable Ultrasound Imaging Sensors for Measuring Fatigue

该项目的目标是开发一种自动辅助装置，能够恢复运动障碍患者的行走和站立功能。虽然对辅助装置的研究，如主动和被动矫形器和外骨骼，已经持续了几十年，但由于一些限制，在行动能力方面的改善一直是适度的。一个主要的挑战是有限的感知和解释人类状态的能力，包括意志运动意图和疲劳。该装置将由电动马达和人体自身肌肉产生的能量组成。这项工作建议开发新的传感器来监测肌肉功能，并且，肌肉疲劳被识别，系统将切换到电动马达，直到肌肉恢复。通过对混合辅助装置的无缝自动控制方法的研究，同时最大限度地减少肌肉疲劳，本研究解决了先前工作的重大局限性。该项目具有长期潜力，可以显著改善脊髓损伤和中风患者的行走和生活质量。拟议的工作还将通过计算算法将基于图像的可穿戴生物传感与物理外骨骼系统集成，从而为网络物理系统的新科学做出贡献。该项目将为研究生和本科生提供沉浸式跨学科培训，将计算方法与成像、机器人、人体功能活动和人工设备相结合，以解决具有挑战性的公共卫生问题。重点将放在让本科生参与研究，作为我们机构结构化课程的一部分。此外，残疾学生将通过利用正在进行的nsf资助项目参与这项研究活动。该项目包括开发可穿戴超声成像传感器和实时图像分析算法，可以直接测量底层肌肉的功能和状态。这将允许开发动态控制分配算法，利用这些信息在驱动和刺激之间分配控制。这种基于肌肉特异性反馈的闭环控制方法代表了传统下肢外骨骼仅依赖关节运动学反馈的范式转变。作为这种新方法的测试平台，该团队将使用一种混合外骨骼，该外骨骼结合了主动关节致动器和对人体自身肌肉的功能性电刺激。重复的电刺激导致肌肉疲劳的快速发作，这限制了这些混合系统的效用，并潜在地增加了受伤的风险。该项目的目标是：开发新型超声传感技术和图像分析算法，用于实时感知肌肉功能和疲劳；研究利用测量肌肉收缩率来最小化疲劳的闭环控制分配算法；将传感和控制方法集成到闭环混合外骨骼系统中，并对脊髓损伤患者进行评估。所提出的方法将通过(1)将人在环物理外骨骼系统与复杂时变物理现象（如动态神经肌肉活动和疲劳）的新型基于图像的实时鲁棒传感相结合，以及(2)开发新的计算模型来解释这些现象并有效地适应控制策略，从而导致创新的CPS科学。这项研究将使实际可穿戴的基于图像的生物传感成为可能，在医疗保健领域有更广泛的应用。该框架可以广泛应用于一些涉及人类参与的医疗CPS问题，包括上肢和下肢假体和外骨骼，康复和手术机器人。基于传感器测量的新型控制分配算法在容错冗余致动器系统中具有更广泛的适用性，可实现无人机的可靠容错控制。

The goal of this project is to develop an automated assistive device capable of restoring walking and standing functions in persons with motor impairments. Although research on assistive devices, such as active and passive orthoses and exoskeletons, has been ongoing for several decades, the improvements in mobility have been modest due to a number of limitations. One major challenge has been the limited ability to sense and interpret the state of the human, including volitional motor intent and fatigue. The proposed device will consist of powered electric motors, as well as the power generated by the person's own muscles. This work proposes to develop novel sensors to monitor muscle function, and, muscle fatigue is identified, the system will switch to the electric motors until the muscles recover. Through research on methods of seamless automated control of a hybrid assistive device while minimizing muscle fatigue, this study addresses significant limitations of prior work. The proposed project has the long-term potential to significantly improve walking and quality of life of individuals with spinal cord injuries and stroke. The proposed work will also contribute to new science of cyber-physical systems by integrating wearable image-based biosensing with physical exoskeleton systems through computational algorithms. This project will provide immersive interdisciplinary training for graduate and undergraduate students to integrate computational methods with imaging, robotics, human functional activity and artificial devices for solving challenging public health problems. A strong emphasis will be placed on involving undergraduate students in research as part of structured programs at our institutions. Additionally, students with disabilities will be involved in this research activities by leveraging an ongoing NSF-funded project. This project includes the development of wearable ultrasound imaging sensors and real-time image analysis algorithms that can provide direct measurement of the function and status of the underlying muscles. This will allow development of dynamic control allocation algorithms that utilize this information to distribute control between actuation and stimulation. This approach for closed-loop control based on muscle-specific feedback represents a paradigm shift from conventional lower extremity exoskeletons that rely only on joint kinematics for feedback. As a testbed for this new approach, the team will utilize a hybrid exoskeleton that combines active joint actuators with functional electrical stimulation of a person's own muscles. Repetitive electrical stimulation leads to the rapid onset of muscle fatigue that limits the utility of these hybrid systems and potentially increases risk of injury. The goals of the project are: develop novel ultrasound sensing technology and image analysis algorithms for real-time sensing of muscle function and fatigue; investigate closed-loop control allocation algorithms utilizing measured muscle contraction rates to minimize fatigue; integrate sensing and control methods into a closed loop hybrid exoskeleton system and evaluate on patients with spinal cord injury. The proposed approach will lead to innovative CPS science by (1) integrating a human-in-the-loop physical exoskeleton system with novel image-based real-time robust sensing of complex time-varying physical phenomena, such as dynamic neuromuscular activity and fatigue, and (2) developing novel computational models to interpret such phenomena and effectively adapt control strategies. This research will enable practical wearable image-based biosensing, with broader applications in healthcare. This framework can be widely applicable in a number of medical CPS problems that involve a human in the loop, including upper and lower extremity prostheses and exoskeletons, rehabilitation and surgical robots. The new control allocation algorithms relying on sensor measurements could have broader applicability in fault-tolerant and redundant actuator systems, and reliable fault-tolerant control of unmanned aerial vehicles.