Research On Multi-mode Control Method For Muscle Strength Training

Rehabilitation treatment of motor dysfunction has been a hot issue in the field of rehabilitation medicine.With the deepening of the basic theory of rehabilitation medicine such as neurophysiology,sports biomechanics and anatomy,the technology of exercise rehabilitation has been greatly enriched and improved.Recovery of muscle strength is the most important part of rehabilitation training.Single-mode rehabilitation training is less efficient.In order to improve efficiency,the combination of multiple training patterns is often used.Based on the analysis of multi-mode rehabilitation training control system elements,the overall architecture of the multi-mode control system is established.Based on the analysis of the muscle mechanics characteristics,the active and passive control algorithms are designed and improved.Based on the request of muscle strength rehabilitation training,the muscle strength rehabilitation experiment system was established,and the man-machine model of upper limb rehabilitation training was analyzed.The multi-mode control scheme for the muscle rehabilitation system was developed.The dynamic model of upper limb single joint was established by dynamic balance method.And the joint dynamics model of multi-joint man-machine is established by Lagrangian method.Based on the general model of muscle force,the elbow joint torque model was established.On the basis of biomechanical model,the mechanism of active and passive rehabilitation training was analyzed.And some active and passive training evaluation indexes were put forward according to clinical experience.According to the requirement of active and passive training,the dual-mode switching control algorithm is designed and the human body model is used to carry on the joint simulation analysis of the controller.To improve the smoothness of the active and passive switching process and reduce the difficulty of switching,the controller structure of the active and passive control algorithm is improved.By establishing a location-based intent recognition algorithm,the accuracy of the active and passive switching is improved.Based on the training requirements and motion characteristics of isokinetic training,the feasibility of multi-joint isokinetic exercise was studied.By comparing the direct matching method with the indirect matching method,the structure of isokinetic training control algorithm is determined.A nonlinear impedance controller is designed on the basis of the impedance controller,and a fuzzy model reference learning control algorithm is proposed.The experimental environment based on dSPACE and embedded controller was set up,and the friction model identification,active and passive control and isokinetic training control experiments were completed.The key points of the control system were verified in order to examine the recovery effect of the control system under the real environment.The experimental results show that the control system is effective.