Research On Iterative Learning Control Of Pneumatic Artificial Muscle-actuated Rehabilitation Robot

In recent years,with the advancement of science and the improvement of the quality of life,people have put forward requirements for medical and rehabilitation methods that are safe,accurate,efficient,and low cost.For hemiplegia and motor dysfunction caused by stroke and limb injury,traditional experience-dependent artificial assistive rehabilitation has been unable to meet these needs.With the characteristics of precise and repeatable trajectory,adjustable safety of auxiliary tasks,and low rehabilitation costs,the emergence of rehabilitation robots perfectly matches the development trend in today's rehabilitation medical field.With the rapid development of rehabilitation robot technology,many challenges have followed,such as the safety of training tasks,the experience of patients during repeated training,and the design of rehabilitation strategies for different rehabilitation stages,all of them have an important influence on the final rehabilitation effect.Targeting the training tasks of rehabilitation robots in different rehabilitation phases of the patient,in order to realize the accurate tracking of the training trajectory in the passive rehabilitation process of rehabilitation robots,and the gradual adjustment of robot compliance and training tasks in the active rehabilitation process,In this paper,the following research work is performed on a parallel rehabilitation robot driven by pneumatic muscle actuator:(1)Aiming at the characteristics of nonlinear and strong modeling of pneumatic artificial muscle,model free adaptive control algorithm based on I/O data driven is studied for the purpose of accurate tracking of training trajectories.Based on this,the model-free adaptive iterative learning control method is studied in order to achieve the progressive tracking of the desired trajectory for the characteristics of repeated movement during rehabilitation training.In order to improve the convergence speed of the algorithm,a model free adaptive iterative learning control algorithm based on high-order parameter estimation is proposed to speed up the iterative estimation of the parameter.(2)For the load disturbance and measurement disturbance caused by the load of the patient's body and the contact friction of the sensor in the passive rehabilitation process.The influences of the disturbance on system output are analyzed by statistical methods.The internal model control principle with disturbance suppression function is studied for the purpose of disturbance suppression.On this basis,a model free adaptive iterative learning control algorithm based on internal model control is proposed,which has a good effect on the disturbance suppression in the passive rehabilitation process.(3)Aiming at the change of the patient's demand for the ideal softness of the robot during the active rehabilitation process,an iterative learning-based impedance control algorithm for of rehabilitation robots is studied.On this basis,the method for assessing the patient's ankle joint exercise capacity is studied,so that the degree of difficulty of the task can be adaptively adjust according to different levels of exercise ability.