Mechanism Design And Control Of Flexible Bionic Tensegrity Ankle Exoskeleton

In recent years,exoskeleton technology has made excellent performance in the fields of helping the elderly and disabled,medical rehabilitation and so on.However,in practical application,the existing exoskeleton has disadvantages such as heavy weight,large size,and difficult to match the degree of freedom of human joint.Therefore,this study develops a flexible bionic tensegrity ankle exoskeleton with light weight and small size based on the design of tensegrity structure,and proposes a data-driven method to control the designed exoskeleton.The main research contents are as follows:(1)Based on the human ankle joint as a prototype,the spatial topology model of the human ankle joint was obtained by analyzing the spatial structure and connection form of the musculoskeletal system near the ankle joint,and the flexible biomimetic tensegrity ankle exoskeleton structure with self-supporting characteristics was designed.The static modeling of the designed exoskeleton is carried out,and the corresponding relationship between the internal force of the string and the balance angle of the exoskeleton is obtained.Based on the principle of minimum potential energy,the optimal cable internal force model is established and solved by quadratic programming method.The results show that the error between the theoretical balance angle and the actual balance angle is less than 0.5°,which verifies the correctness of the model.(2)Combined with the traditional model-based tensegrity control method to control the designed ankle exoskeleton,explored the trajectory tracking performance and robustness of the ankle exoskeleton under the power assisted control methods based on the position model,the statics model and the force position mixed model.The experimental results show that the position model control method has good tracking performance and robustness,and can be used for flexibility training in early rehabilitation training.(3)In order to make the ankle exoskeleton suitable for more complex rehabilitation scenarios,a data-driven control strategy was proposed.The strategy trained the reinforcement learning network through the data obtained in the experiment to obtain the corresponding control model,which was used to realize the control of the designed ankle exoskeleton.The experimental results show that the target angle trajectory of the ankle joint with a period of 1.22 s can be tracked under the data-driven control mode,and the average error is less than 0.5° and the peak error is less than 1°,which verifies the feasibility of the data-driven control strategy.(4)With the rehabilitation performance of the designed flexible biomimetic tensegrity ankle exoskeleton as the research objective,high-speed cameras and electromyographic sensors were used to collect human gait data during walking.The experimental results showed that after wearing the ankle exoskeleton,The RMS of major muscle electrical signal activity decreased by 17.8%,11.8% and 9.5% under three different gaits,which verified the effectiveness of ankle exoskeleton assisted rehabilitation.