| Stroke disease has characteristics of high disability rate,youthful trend and great harm,which seriously endangers the physical and psychological health of patients and brings heavy economic burden to their families and society.For this reason,the researchers have designed a variety of upper limb rehabilitation mechanisms,among them,the wear upper limb rehabilitation robot has become a hotspot in the rehabilitation medicine field due to its strong flexibility and high rehabilitation efficiency.The wear upper limb rehabilitation robot is a rehabilitation device,which is designed to assist patients with upper limb movement disorders specifically.However,the current research on wear upper limb rehabilitation robot systems still encounters the problems which are included the complex mechanical structure,weak intention recognition ability,and sharp decline in human-machine interaction control performance under complex working conditions.Therefore,to solve the above problems,this paper focuses on the structural design of the wear upper limb rehabilitation robot,the active movement intention recognition,and the human-machine interaction control method.The main content and innovative results of this paper are proposed as follows:(1)The structural design and dynamics analysis problems of upper limb rehabilitation robots are proposed and analyzed in this section.Analyzing the upper limb skeletal muscle anatomy and upper limb movement characteristics,which are simplified the upper limb structure into a connecting strut and hinge structure,so that the corresponding skeletal structure can rotate around three mutually perpendicular coordinate axes,and the rotation center conforms to the biology principles of human joints movement.In addition,the multi-joint movement mechanism of the human upper limb is analyzed and combined with human anatomy to build a wear upper limb rehabilitation robot that considers the shoulder joint,elbow joint and wrist joint.For reducing the complexity of the structure,a combination of rope drive and motor drive are utilized to optimize the size and weight of the mechanical structure,which can improve the flexibility and controllability of the upper limb rehabilitation robot.The dynamics analysis of the designed wear upper limb rehabilitation robot is carried out which is based on the Lagrange method.It may lay the foundation for the design of humanmachine interaction control method.(2)The identifying active movement intentions of multiple joints of the upper limbs is developed in this section.The s EMG signals of the relevant muscles of the human upper limbs are collected,and an open-loop prediction model is designed which is based on the radial basis function neural network and the long short-term memory neural network.It can provide an accurate identification solution for the prediction of continuous motion of multiple joints of the human upper limbs.Besides,to solve various noise interference problems with s EMG signals in clinical applications,this paper designs a closed-loop prediction based on the long-short-term memory neural network and a zeroing neural network with noise suppression capabilities,which can improve the performance of the multiple joints of upper limbs under noise interference conditions.The performance of continuous motion estimation can realize the effective information interaction between the human body and the upper limb rehabilitation robot.Finally,experimental results infer that the closed-loop model has extremely strong stability and high generalization ability,which can provide a guarantee for the establishment of a comfortable and safe human-machine interaction control system.(3)The human-machine interaction control method of the upper limb rehabilitation robot is designed in this section.Based on the dynamics model of the wear upper limb rehabilitation robot and the active movement intention of human upper limb multi-joints,a zeroing neural network controller with exponential convergence is designed to drive the wear upper limb rehabilitation robot to follow the active movement intention of the human body to achieve trajectory tracking control.Furthermore,for the unavoidable existence of mechanical friction,model perturbation,parameter uptake and other disturbances in the rehabilitation process,an integral control term is introduced on the basis of the zeroing neural dynamic model,and then a noise-suppressing zeroing neural network controller(NSZNN)is designed,which can suppress different external disturbances and ensure that the rehabilitation training process is not affected by the noise and disturbances.The experimental results show that the NSZNN controller proposed in this paper is able to complete the trajectory tracking control based on the active motion intention of the human body under different disturbances,and has stronger convergence performance and robustness compared with the traditional controller.(4)The experimental platform system of wear upper limb rehabilitation robot is designed,and the upper limb rehabilitation training experimental research are proposed.The experimental platform of wear upper limb rehabilitation robot is built to meet the demand of upper limb rehabilitation training of subjects.Furthermore,the unloaded experiment of wear upper limb rehabilitation robot is carried out by using proportionalderivative(PD)control algorithm to verify the tracking performance and dynamic response performance of the wear upper limb rehabilitation robot and ensure the safety of human-machine interaction system.Besides,the human-machine co-operative control method using NSZNN controller is exploited to carry out the experimental procedure of upper limb rehabilitation training for the wear upper limb rehabilitation robot system.The single-joint and multi-joint experiments can prove the superior tracking performance of the human-machine cooperative control system of the wear upper limb rehabilitation robot which is designed in this paper.It can achieve the single-joint and multi-joint rehabilitation training activities in different modes. |
PDF Full Text Request