【作者】 陈强；

【导师】 金弘哲；

【作者基本信息】 哈尔滨工业大学 ， 机械电子工程， 2013， 硕士

【摘要】 病理性震颤是一种常见于上肢的疾病，影响着患者的生活质量和心理健康。随着技术的发展，对于病理性震颤的治疗也逐渐有所成效，其中外骨骼辅助治疗手段是近几年发展起来的治疗方式，在生物医学领域具有广阔的应用前景。新一代抑震机器人将采用功能性电刺激作为抑震手段，本次研究的主要目的是建立一个适用于功能性电刺激抑震的分析模型。本文基于肌肉的Hill模型的理论，探究了肌肉电信号与肌肉活化度间的转换关系，并利用开放性生物力学仿真软件OpenSim建立了人体上肢骨肌系统模型，最终实现了以肌肉活化度为驱动的骨肌系统模型。肌肉活化度是一个表征肌肉兴奋状态的抽象概念，而表面肌电信号(sEMG)也能反映肌肉的活跃程度，因此两者存在着一定的联系。为将真实的震颤信息输入骨肌系统模型，需将易测量的表面肌电信号转化为模型的肌肉活化度，因此，设计了一个新的计算方法——Hill模型与神经网络相结合的方法，实现了表面肌电信号到肌肉活化度的转换，经过转换后的表面肌电信号可驱动建立的骨肌模型，并能模拟肢体的震颤运动。本文对震颤运动也进行了初步研究，应用经验模态分解方法(EmpiricalMode Decomposition，简称EMD)对震颤患者的关节运动信号和肌电信号进行了分析，实现了意向运动信号和震颤运动信号的分离以及信号的时频域分析。提出了判断震颤发生的时域指标和频域指标，有效地对震颤运动进行了识别，并可用于抑震机器人对震颤发生的判断。在实验环节，基于骨肌系统模型和表面肌电信号对震颤运动进行了仿真，分析了骨肌模型对输入肌肉活化度的频率响应特性，分析了影响上肢频率响应的主要因素，建立了肌电信号与关节力矩之间的对应关系模型，并利用静态优化算法对肌肉力的分配进行了优化处理。更多还原

【Abstract】 Pathological tremor is a disease and usually appears in upper limb, while thequality of life and mental health of patients are interfered. With the developedtechnology, the treatments of pathological tremor are also gradually improved. Inrecent years, exoskeleton adjuvant therapy is developed as a new treatment and hasbroad application prospects in the biomedical field. Functional electrical stimulationmethods will be employed for the next-generation tremor suppression robots. Dueto the requirement of simulation platform for tremor suppression, the biomechanicalmodel of upper limb was studied in this paper.Based on the theory of the Hill muscle model, the paper explores therelationship between the surface electromyography (sEMG) and the degree ofmuscle activation. A musculoskeletal system model of upper limb was ultimatelyestablished by the biomechanical simulation software-OpenSim and this model wasdriven by the degree of muscle activation.Muscle activity is an abstract concept used to characterize the degree of muscleactivity and the sEMG of muscles can also objectively reflect the level of muscleactivity, therefore the muscle activity is related with the sEMG. The Hill model ofmuscle was employed with neural network to convert the sEMG into the musclesactivity. Finally, the musculoskeletal model was active by the converted sEMG andthe model can also simulate the tremor movement of limb.The tremor movements are primarily analyzed in this article. The sEMG and theelbow joint kinematic signals of patients with pathological tremor are analyzed byusing the empirical mode decomposition method (EMD). The intentional movementsignals and tremor signals are separated and time-frequency analysis of signals arecompleted by the EMD. The indicators of time domain and frequency domain areproposed for discriminating the tremor from subjects’ normal activities and this isalso employed for the tremor suppression robot to evaluate the tremor.In the experiments, the musculoskeletal system is used to simulate the tremormovement. Using the simulation results, the frequency response characteristics ofthe model with muscle activity as the input and the factors impacting the frequencyresponse of limb are both analyzed. The relationship between EMG and joint torqueare identified based on the experimental data. Finally, the optimal distribution ofmuscle forces is completed by the static optimization algorithm.更多还原