导　　师： 王建辉

学科专业： 081101

授予学位： 硕士

作　　者： ;

机构地区： 东北大学

摘　　要： 上肢康复机器人是近几年来迅速发展的一门新兴技术,是机器人技术在医学领域的新应用,利用康复机器人对患者进行康复训练可以提高偏瘫康复训练的针对性及科学性,能为患者制定更好的康复方案,进一步提高康复的效率。然而,康复机器人本身是一个具有很强的非线性、强耦合的系统,为了实现对机器人期望轨迹的高精度跟踪,本文以五自由度上肢康复机器人作为控制对象,研究迭代学习控制算法在其上的应用。 首先,本文在坐标变换理论的基础上,对上肢康复机器人的运动学系统作了较深入的分析与研究,并且应用Matlab仿真软件中的SimMechanics和虚拟现实工具箱,按照实际机器人参数,建立了上肢康复机器人的运动学仿真模型,并对其进行了仿真,结果表明,三维模型的仿真效果更直观,可以更清楚、更真实的反应康复机器人手臂的动态运动情况。 然后,通过对上肢康复机器人机械结构特点及运动特性的分析,将机器人简化为一个空间三自由度机械手,利用拉格朗日方程法和牛顿-欧拉方程方法分别求解了三个关节的广义驱动力矩的表达式,并且对两种方法进行了对比,为上肢康复机器人的实时控制和动态仿真等复杂问题的研究奠定了理论基础。 最后,本文主要介绍了迭代学习控制的基本概念,包括开环迭代学习控制和闭环迭代学习控制,分别对其收敛条件进行了分析,并且讨论了其在上肢康复机器人上的应用。在此基础上对闭环学习算法作了改进,给出了指数变增益加速学习算法,并在上肢康复机器人动力学模型上进行了仿真验证,该算法可以大大提高学习速度,保证了系统的快速跟踪,起到了加速学习的作用。 Upper limb rehabilitation robot is a new technology that has been developed rapidly in recent years. Using the rehabilitation robot on patients with rehabilitation training can improve the relevance and scientific of hemiplegia rehabilitation, can formulate better rehabilitation programs for patients, and further improve the efficiency of rehabilitation. However, rehabilitation robot itself is a highly nonlinear and strong coupling system. In order to achieve the desired trajectory of robot high accuracy tracking, the work of this thesis is to research iterative learning control and its application for the5-DOF upper limb rehabilitation robot. First of all, this thesis does much research on upper limb rehabilitation robot kinematics system based on coordinate transformation theory. By using SimMechanics toolbox and virtual reality toolbox in Matlab simulation software, kinematics simulation models of upper limb rehabilitation robot is established by the actual parameters of robot. The results show that the3D model simulation effect is more intuitive, by which rehabilitation robots dynamic movement can be more clearly and more really reacted. Then, by analyzing the structure and motion characteristics of the upper limb rehabilitation robot, the5-DOF upper limb rehabilitation robot is simplified to a space three degrees of freedom mechanical arm. The analytical solutions of three generalized joint torques are solved by Lagrange equation and Newton-Euler equation, the results are cleaned up and compared, which is provided as a theoretical basis for the research on real-time control and dynamic simulation. Finally, this thesis mainly introduces the basic concept of iterative learning control, including the open-loop iterative learning control and closed-loop iterative learning control. The convergence conditions for them are discussed separately. And also discuss their application on the5-DOF upper limb rehabilitation robot. On the basis of that, closed-loop learning algorithm is improved. An index variab

关 键 词： 上肢康复机器人 迭代学习 运动学 动力学

领　　域： [自动化与计算机技术] [自动化与计算机技术] [自动化与计算机技术] [自动化与计算机技术]