Upper Limb Rehabilitation Robot Iterative Learning Control

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 variable gain accelerated learning algorithm is presented. This algorithm can greatly improve the iterative learning speed, and can ensure the fast tracking of the system at the same time, which plays the role of accelerated learning. In the upper limb rehabilitation robot dynamics model these are verified through simulation test analysis.