Hybrid parallel compliant actuation for lower limb rehabilitation

麻痹是骨骼肌组织中单个或多个肌肉的功能丧失。针对瘫痪的原因，有多种治疗方法。一方面，可以通过自动运动治疗实现康复训练，另一方面，可以通过动力执行器替代运动功能缺失的补偿。本课题将从根本上研究一类新的外骨骼致动器和控制装置以及主动矫形器，特别适用于残功能患者（偏瘫、中风幸存者）。截瘫患者的适用性将在项目的后期阶段进行检查。作为核心部件，本项目重点设计用于下肢运动支撑的新型可变平行弹性致动器（VPEA）。为了能够完全控制进入并联弹簧的能量流，执行器配备了制动器和离合器。这种方法旨在最大限度地减少必要的致动器功耗，并在治疗期间使具有残余运动功能的患者更深入地参与运动。在作动器设计的同时，还需要研究非线性动力学模型，该模型除了包含混合机电模型外，还包含整个系统的生理和动力学子模型。执行器复杂性的增加要求多目标控制设计。除了级联力和阻抗控制器（根据需要辅助原则）外，必须控制流入平行弹性元件的能量。此外，必须针对具有多个驱动关节的过驱动系统进行控制综合设计。此外，控制器必须保证与其佩戴者可能的残余肌肉功能相关的交互稳定性。为了能够实现这些目标，研究了一种具有切换系统鲁棒控制、迭代学习控制或模型预测控制可能组合的级联控制结构。我们的目标是创建一个经过验证的康复系统，可以用于这两个患者群体。在将驱动器集成到现有的外骨骼结构和运动测试中之后，还将结合VPEA对商用FES系统进行测试。为此，需要开发一种协同的生理控制器，它可以使用患者的肌肉和VPEA一起进行运动。该工作包的总体目标是将外骨骼的适用性扩展到截瘫患者群体。执行器，包括单个关节的底层控制系统，依次投入运行并进行测试。最后，整个外骨骼支持系统在北京清华大学中国运动实验室的跑步机研究中得到验证，该实验室从一家合作的中国医院招募了交叉患者。

Paralysis is a functional loss of a single or multiple muscles of the skeletal musculature. Corresponding to the cause of the paralysis various therapy approaches are possible. On the one hand, a rehabilitative training can be realised with automated motion therapy and, on the other hand, the compensation of missing motor function can be substituted by powered actuators. In this research project, a new class of actuators and control of exoskeletons and active orthosis are to be fundamentally researched, which are especially suitable for patients with residual function (hemiplegics, stroke survivors). The suitability for paraplegic patients is to be examined in a later phase of the project. As a core component, the project focuses on the design of novel variable-parallel-elastic actuators (VPEA) for the movement support of the lower limbs. In order to be able to fully control the energy flow into the parallel spring, the actuator is equipped with a brake and a clutch. This approach is intended to minimise the necessary actuator power consumption and to involve the patient with a residual movement function more intensively in the motion during therapy due. In parallel to the design of the actuator, a nonlinear dynamical model is to be researched, which also contains physiological and dynamic submodels of the overall system, in addition to the hybrid electromechanical model. The increased complexity of the actuator demands for a multiobjective control design. In addition to cascaded force and impedance controllers (assist as needed principle), the energy flow into the parallel elastic element has to be controlled. Moreover, the control synthesis has to be designed with respect to an overactuated system with multiple actuated joints. Furthermore, the controller has to guarantee interaction stability with respect to possible residual muscle function of its wearer. To be able to achieve these objectives, a cascaded control structure with possible combinations of robust control for switched systems, iterative learning control or model-predictive control are investigated.The goal is to create a proven rehabilitation system that can be used for both patient groups. After integration of the actuator into an existing exoskeleton structure and motion tests, a commercial FES system is also examined in combination with VPEA. For this purpose, a synergistic physiological controller is to be developed which uses the patient muscles and VPEA together for the movement. The overall goal of this work package is the extension of the exoskeleton applicability to the group of paraplegic patients.The actuators, including the underlying control system for the individual joints, are successively put into operation and tested. Finally the whole exoskeletal support system is validated in a treadmill study in a Chinese motion lab of the Tsinghua University of Peking with cross-parlayed patients recruited from a cooperating Chinese hospital.