人手自然运动机理及其在假肢手运动生成中的应用

It is a key fundamental problem how to model the dynamic movement of hand in the grasping and manipulating, reveal the motion principle of the hand synergy and apply to the motion generation of humanoid prosthetic hand. On this problem, the sampled data of hand movement from motion capture experiment is used to analyze the coordinated mechanism of the temporal postural synergy, and the kinematic model of the temporal postural synergy is proposed. Based on the biomechanical model of hand and the method of surface electromyography, the motor synergies in the form of angular acceleration, joint torque and muscle activation are analyzed. Using the analysis result, the computational model of dynamic synergy is proposed. The movement raised by the principal components with un-significant variation at the total variance of motion, called secondary movement, is also analyzed to explore its dynamic principle. The model about the dynamic principle of secondary movement of human hand is also proposed. Based on this model, the dynamic synergy of prosthetic hand describing the evolvement from primary pattern movement to the secondary movement is developed. In this model, the parameters representing the joint torque and tactile information of prosthetic hand in manipulation are included. In order to generate the posture trajectory of prosthetic hand similar to that of original human hand as much as possible, the nonlinear mapping relationship among the axis components of the trajectory in the eigengrasp space is studied. This nonlinear mapping relationship is used to transform the super-curve in high dimension space into the 3D curve in low dimension space. Based above developed techniques and strategies, we try to establish the systematic theories and methods to implement the flexible movement of prosthetic hand similar to the human hand, for the destination of accumulating critical techniques and common fundamental theories for developing the practicable biomechatronic prosthetic hand in clinic environment.

针对如何描述人手在日常抓取与操作中的动态运动行为，并将人手的自然运动生成机理应用到肌电假肢手的拟人运动规划与控制这一关键基础科学问题，结合人手运动解析实验，分析人手手势随时间的协调变化规律，提出时变手势协同的运动学描述方法。建立人手肌骨的生物力学模型，结合表面机电信号，分析人手关节角加速度、关节力矩以及相关肌群的协调机制，提出人手动态协同描述方法。分析人手次运动的动力学机理，建立次运动的动态协同模型，利用假肢手的力觉和触觉传感信息，建立假肢手从主模式运动演化为次运动的动态协同主元模型。研究特征抓取子空间内运动轨迹各坐标分量之间的非线性映射关系，提出与生机接口功能匹配的人手动态手势高保真复现方法。在此基础上，揭示手指动态协同的运动机理，提出能够在生机接口支配下实现假肢手灵巧操作的理论与方法，为具有实际临床应用价值的灵活假肢手的研发和应用积累关键技术和共性基础科学理论。

如何工程科学方法修复手臂缺失患者的运动功能，实现假肢手在患者自主意识下灵活运动，是目前极具研究希望，可望实现运动功能缺失患者肢体“再生”的有效途径。针对这一挑战性问题，本项目围绕人手高维抓握运动的降维与重构，自然运动时空特征建模及其运动生成，主次运动分治的假肢手驱动机构设计等内容开展系统深入的基础研究，揭示了人手自然抓握运动的生成规律，建立了欠驱动拟人假肢手设计与运动控制的理论与方法，为实现拟人假肢手的人体运动意图控制提供理论支撑。. 结合日常生活常见的抓取模式，构造了蕴含人手抓握运动信息的高维运动数据集，建立了人手特征速度的提取方法，提出了特征运动速度的机械实现原理，建立了通过特征凸轮轴来实现全局耦合的假肢手机构设计理论。针对复现人肢体自然运动时空特征的问题，提出了可近似描述人肢体运动固有时空特征的二阶运动学建模方法，建立了人手空间位置与肢体姿态坐标的映射关系，提出了人体上肢等效运动链的运动学反解方法，为臂-手假肢实现拟人功能运动提供理论支撑。研究了人手指间协同以及指内的协同规律，建立了人手协同运动基植入拟人假肢手的机械实现原理，实现了通过低维机械运动复现高维抓握运动的方法，提出了紧凑的剪式欠驱动机构设计原理，解决了假肢手机-电集成的难题。. 以上研究结果加深了对人手运动机理的认识，创新了假肢手机构设计方法，为生机电一体华假肢手的临床应用提供了技术积累与理论支撑，拓展了神经接入假肢手的控制方法与理论，对带动制造科学与神经工程学、信息科学交叉学科的发展，促进我国康复工程技术的进步具有重要的意义。