Upper Limb Subject-Specific Modeling And Muscle Forces Predicting Based On Matlab-OpenSim Software Interface

The rehabilitation robot is becoming the main physical therapy for human hemiplegia.Biofeedback in rehabilitation robot is a hot research issue.Focusing on biological information feedback in rehabilitation robot,this paper established subjectspecific musculoskeletal model and muscle force prediction model to predict individual muscle force including biceps long and short head,triceps lateral,medial and long head,brachialis,brachioradialis and elbow joint torque;Based on motion capture and EMG signal synchronization platform,the subject-specific model was established.Through subject-specific model,muscle fiber lengths and muscle joint moment arms were obtained.The forward dynamics method was used to obtain the muscle activation and muscle force.The modeling method of upper limb musculoskeletal model,the processing method of EMG signal to muscle activation,and the mathematical characterization method of muscle physiologic movement were systematically discussed.The main contents and achievements are as follows:(1)A subject-specific upper limb musculoskeletal model was established for individual differences.Firstly,the mass,length,center of mass and moment of inertia of each link of the upper limb were calculated through anthropometry.Using static motion capture technology,the local coordinate system model of upper limb joint was constructed by vector method.Using open-source software Open Sim and its visual bone database,skeletal muscle model was constructed by XML programming.The inverse dynamics analysis of the model was carried out,and the influence of the model on the joint torque was analyzed by comparing the results of literatures.The results show that the accurate kinematics,dynamics,muscle fiber length and torque data can be obtained by the subjectspecific musculoskeletal model of the upper limb.(2)The muscle activation dynamics model was established,and coded in MATLAB to realize the retrieval of the activation signals transmitted to the muscles from nerves.Firstly,the EMG signal and the maximum voluntary contraction signal were collected.And the characteristics of EMG signal are analyzed by Fourier transform.The filter is designed by analyzing the frequency domain signal.A muscle activation model dynamics was established to simulate the degree of muscle activation in the process of nerve impulse conduction to muscle generating muscle force.The results showed that the muscle activation parameters representing the muscle activity parameters could be obtained from the muscle activation dynamics model.(3)The muscle force prediction model was established,and the model parameters were optimized by simulated annealing algorithm.Firstly,the muscle tendon components were modeled mathematically.The modified Gaussian function was used to describe the relationship between muscle force and muscle fiber length,the logarithmic function was used to describe the relationship between muscle force and velocity,and the piecewise function was used to describe the relationship between tendon stress and strain.Secondly,the relationship between the muscle force and the pinnate angle was discussed,and the effect of the pinnacle on the muscle force was introduced into the muscle prediction model.Through the Open Sim API,the muscle fiber lengths and muscle joint arms of the model were calculated,so the joint torque was obtained.Finally,the model parameters were optimized by simulated annealing algorithm.In this paper,motion capture and EMG signal synchronous acquisition experiments were carried out to verify the correctness of the muscle force prediction model.The Pearson coefficient between the results and the measured joint torques was 0.91,showing a strong correlation.And the deviation between the results and the measured joint torques was 0.35 within a reasonable range.Thus,the muscle force prediction model can be used to predict muscle force and joint torque,which can be used to provide biological information feedback for rehabilitation robots.It is of great significance to realize the intelligent rehabilitation equipment,intelligent diagnosis and quantification.