Research On Active Training Method Of Upper Extremity Rehabilitation Exoskeleton Based On Admittance Control

The upper extremity rehabilitation exoskeleton is a kind of medical device that assists patients with upper limb motor dysfunction in rehabilitation training.In recent years,due to neurological diseases and other problems,the demand for upper limb rehabilitation has increased and upper extremity rehabilitation exoskeleton has attracted much attention.The use of upper extremity exoskeleton to provide patients with efficient and intelligent rehabilitation training is of great significance to make up for the shortcomings of traditional rehabilitation and improve rehabilitation efficiency.Therefore,based on the upper extremity exoskeleton,this paper aims to provide patients with more appropriate and effective active training.The research mainly focuses on the two key issues of patient's active intention detection and design of active training methods.Accurately obtaining the patient's active motion intention is a necessary part of implementing active training.Designing the corresponding auxiliary method according to the degree of the patient's movement disorder is the core link to implement active training.In view of the lack of accuracy,stability and completeness in the detection of current patient's active motion intentions,in this paper,based on the analysis of the human-machine physical interaction of the upper extremity exoskeleton and the feasibility requirements of engineering applications,we decided to use the human-machine interaction torque as a quantitative indicator of active motion intention,and studied two methods based on contact force and inverse dynamic model to obtain the human-machine interaction torque.The two methods are essentially the same,but differ in the way they are implemented.In terms of theory,they depend on the kinematics model and the inverse dynamics model,respectively.In terms of sensor data,they depend on human-machine contact force and joint driving torque,respectively.In this paper,the theoretical and simulation experiments of these two methods are analyzed.The results verify that the method based on contact force is simpler and more stable but there is a certain lack of completeness,and the method based on inverse dynamics is more complete but subject to the accuracy of the model in terms of accuracy.In view of the problem that the current active training does not fully match the degree of patient's movement disorder,this paper uses admittance control framework to design cooperative active training and autonomous active training based on human-machine interaction torque.The core idea of admittance control is to convert human-machine interaction torque into position offset.Cooperative active training divides the human-machine dynamic interaction area based on the terminal posture error of the upper extremity exoskeleton,so as to adjust the correction effect of position offset on the predetermined track,and then ensure the active participation of patients under the traction and constraint of the predetermined track.Autonomous active training assists patients to perform autonomous movements under the guidance of virtual games through the way of tracking the position offset by upper extremity exoskeleton,so as to ensure that the patients occupy a dominant position in the movement process.In this paper,the principle analysis of these two active trainings is carried out and the method is verified by Simulink and Adams joint simulation experiments.In the simulation results,the degree of human-machine movement consistency and the interaction force indicate that these two active training methods meet the requirements of patients with different degrees of movement disorder for human-machine interaction and rehabilitation training.In the end,based on the built upper extremity exoskeleton platform,the functional tests of these two active training methods were carried out,so as to verify the feasibility of the method of simulation study in the actual system.