Study On Upper Limb Exoskeleton Robot Rehabilitation Training System

The rehabilitation training system of exoskeleton robots are developed for upper limb motor dysfunction caused by stroke sequelae in this dissertation.The theory of neural plasticity,facilitation technology and exercise relearning method,combined with traditional rehabilitation medicine,human anato my,rehabilitation engineering and robotics are proposed by using information fusion,intelligent computer technology.The medical staff can be replaced by the exoskeleton to help patients with relavant rehabilitation training.Furthemore,the exoskeleton can maintain joint mobility,enhance muscle strength and improve the coordination of limbs and muscles,to a certain extent,stimulate the damaged nerve to achieve self-repair.An adjustable exoskeleton of 6 DOFs modular parallel drive joint is designed from the point of view of safety,effectiveness,practicability and comfort of clinical rehabilitation process on the basis of the mechanism of upper limb joint movement and the upper limb anatomy of the human body.It is suitable for training of human upper limb single joint or polyarticular rehabilitation exercise in sitting or standing condition.The modular structure of the exoskeleton makes the rehabilitation of single joints be more convenient and flexible.In addition,the design and application of the spatial gravity balance mechanism based on the law of conservation of energy have realized the self-balance of most of the exoskeleton itself,which ensured the safety of the rehabilitation process and reduced the driving cost to a certain extent.Based on the generation mechanism of SEMG,the probabilistic relationship of regularity and randomness between SEMG and joint angle was proposed.The principal component analysis method is used to reduce the dimensionality of SEMG,and the SEMG-angle HMM model,which can be used to construct SEMG-joint angle library off-line for predicting the joint angles.Compared with the BP neural network and RBF neural network modeling,HMM modeling which can achieve higher accuracy and better online prediction is adopted in this dissertation.SEMG can represent subjective initiative of huamn movement.The application of SEMG in the rehabilitation system can stimulate the rehabilitation of subject,accelerate the remodeling of the nervous system and improve the efficiency of rehabilitation.The interactive control strategy of the combination of SEMG and TLC includes joint torque loop and error regulator.Where,the torque loop is based on the dynamic model of the exoskeleton and the error regulator which can adjust the error online in real time is based on PD spectrum theory.The experimental results shown that the controller had a good trajectory tracking,anti-jamming performance and robustness.According to the different rehabilitation stages of the subjects,passive and active rehabilitation training mode are designed.Single joint single degree of freedom exercise training and multi-joint multi-degree of freedom compound exercise training are carried out on the basis of traditional rehabilitation training method,in each rehabilitation training mode.Finally,the experimental platform of upper limb exoskeleton rehabilitation training system is set up.After the relevant experimental study,the analysis of experimental results proved that the system can realize the rehabilitation training of single joint/polyarthhesis in different modes.The SEMG and TLC-based controller can meet the requirements of clinical rehabilitation training in active rehabilitation training and achieve a goo d trajectory tracking effect.