SCH: Spine-Hip Exoskeletons with Learning-Based Optimal Control for Low Back Pain Alleviation

这一合作项目将利用脊柱-髋关节外骨骼，一种跨学科和综合性的 联合快速发展的科学和工程领域的平台，以促进知识和 在其领域内和跨不同领域的理解。最先进的外骨骼缺乏人类 减轻工人伤害的适应性和任务通用性。此外，人体背部的解剖 对可穿戴机器人的设计和控制提出了独特的挑战。因此，脊椎外骨骼是 需要减少至少三种力中的一种(也不增加其他力)，包括竖起棘肌 肌力和腰椎的压缩和剪力。这就需要新的解决方案 机器人的设计、建模和控制以实现所有目标。这个项目将1)发展 力学引导的脊柱-臀部软外骨骼，2)了解高保真的肌肉骨骼模型 人类脊柱及其对外骨骼的反应，以及3)研究基于学习的最优控制 减少肌肉骨骼关节的负荷，从而最终减轻工人的腰部损伤。我们的 多学科团队由机器人专家(苏浩博士)、计算生物力学专家(苏浩博士)组成。 和基于学习的最优控制(江中平博士)将采取收敛 使用由脊柱组成的仿生动力外骨骼辅助多个关节的方法 和臀部模块，用于防止执行举重任务的工人的下腰痛(包括下蹲和 弯腰姿势)。

This collaborative project will leverage the spine-hip exoskeleton, an interdisciplinary and integrative platform uniting rapidly advancing areas of science and engineering to advance knowledge and understanding within its field and across different fields. The state-of-the-art exoskeletons lack human adaptability and task versatility for injury mitigation of workers. Moreover, the anatomy of the human back presents unique challenges for the design and control of wearable robots. Thus spine exoskeletons are required to reduce at least one of three forces (also not increase other forces), including erector spinae muscle force and lumbar vertebral compressive and shear forces. This necessitates new solutions for robot design, modeling, and control to achieve all objectives. This project will 1) develop mechanics-guided spine-hip soft exoskeletons, 2) understand the high-fidelity musculoskeletal model of the human spine and its response to exoskeletons, and 3) investigate learning-based optimal control to reduce musculoskeletal joint loadings thus ultimately mitigate low back injuries to workers. Our multidisciplinary team consisting of experts in robotics (Dr. Hao Su), computational biomechanics (Dr. Katherine Saul), and learning-based optimal control (Dr. Zhong-Ping Jiang) will take a convergent approach to assist multiple joints using a bio-inspired powered soft exoskeleton composed of the spine and hip modules for low back pain prevention of workers who conduct lifting tasks (including squat and stoop postures).