Mechanical Design Of Hand Rehabilitation Robot And Research On Human-machine Interaction Strategy

Along with the advancement of technology,people’s pursuit of high-quality medical care continues to rise,and rehabilitation robots,as an emerging technology,are a boon to many stroke patients.As research progresses,rehabilitation training for patients is no longer limited to the major joints of the extremities;the more dexterous movement of the hands has also become a current focus.At present,the hand rehabilitation robot is still deficient in fine motor function,cooperative control method,interactive training mode,and motion state monitoring.Focusing on the above-mentioned issues,this thesis conducts rehabilitation system design and key technology research of hand rehabilitation robot from five aspects: configuration design and motion analysis,compliance control and cooperative motion,human-machine interaction and paradigm design,intention recognition and fatigue determination,platform construction and experimental verification.First,based on the hand anatomy and joint motion mechanism,the motion coupling characteristics are analyzed and the equivalent mechanism model of the hand is established.Aiming at the adduction/abduction movement which is often neglected in hand rehabilitation training,study of the relationship between flexion/extension motion and adduction/abduction motion is conducted.On this basis,the mechanism of hand rehabilitation robot is designed according to the design requirements and the motion characteristics of fingers.Based on the concept of lightweight design,a remote drive structure based on the Bowden cable is proposed to reduce the burden on the hands during training.Based on the previous research,the prototype of the hand rehabilitation robot system is completed.The kinematic simulation and analysis of the proposed mechanism are performed with the index finger as an example,and the workspace of the robot is solved.Next,to ensure the accuracy and safety in the rehabilitation training process,the control strategy of hand rehabilitation robot is carried out.Considering the error characteristics of remote drive,a sliding model control strategy based on output delay observer is designed for robot position control.To ensure the flexibility of rehabilitation training,a position-based impedance control strategy is designed.Based on the clinical therapy of isometric training and isokinetic training,a trigger assisted control strategy combining active and passive is proposed.To solve the synchronization problem in the process of multi-finger cooperative movement,a cross-coupling synchronous control strategy is designed and simulated.Then,to realize the effective combination of motor rehabilitation and neurological rehabilitation in the training process of patients,the research of human-machine interaction rehabilitation training model is carried out.Firstly,the hand rehabilitation mode is analyzed based on the clinical requirements,and a comprehensive training program combining motor rehabilitation and neurological rehabilitation is designed under the virtual reality scene,and the effectiveness of the program is verified by the eye-movement information during the rehabilitation process.To further improve the effectiveness of neurorehabilitation training,a visual stimulation-verbal enhancement hand rehabilitation training paradigm is proposed,and the experiments are designed to verify the effectiveness of the paradigm in the neurorehabilitation training process by studying the response of the cerebral cortex under multi-sensory stimulation.After that,to improve the intelligence of hand rehabilitation training,a study on motor intention recognition and fatigue determination based on s EMG signals is conducted.The s EMG signals of four volunteers are collected and subjected to preprocessing,feature optimization and extraction.The obtained feature matrix is input into LSSVM for motion intention recognition of four hand movements.An improved arithmetic optimization algorithm is proposed to optimize the hyperparameters in the LSSVM to improve the generalization ability of the classifier.A labeling classification method for patient fatigue including transition states is proposed,and the k-NN method is used to determine the fatigue levels of volunteers.Finally,the experimental research will be conducted around the built hand rehabilitation robot system.Firstly,the remote drive transmission efficiency and joint motion angle of the hand rehabilitation robot are tested.Then experiments are conducted to verify the grasping ability,stability and adaptability of the robot.Then the force feedback effects of the active grip scenario are tested.After that,6 volunteers are recruited to test the algorithm of hand motion intention recognition and fatigue determination based on s EMG.Finally,with the participation of 25 healthy volunteers,the fuzzy analytic hierarchy process is used to pre-evaluate the comfort of the robot system.The results show that the hand rehabilitation robot system designed in this thesis could meet the rehabilitation requirements of patients.