Design And Control Method Of Bionic Knee Exoskeleton Based On Tensegrity Structure

With the rapid growth of the aging population in China,the number of physical disabilities caused by stroke and other diseases is increasing rapidly.Therefore,the lower limb rehabilitation machine has entered the golden period of development and has a broad market prospect,however,traditional postoperative rehabilitation methods have disadvantages such as insufficient number of rehabilitation therapists,long rehabilitation cycle and large capital consumption.In the process of stoke postoperative rehabilitation,due to the active motion intention of the affected limb,the actual motion trajectory of the affected limb will be mismatched with the rehabilitation robot,which is easy to cause second injury to the affected limb.In addition,the lower limb rehabilitation robot has disadvantages such as complex structure,insufficient bionic performance and poor anti-noise ability of the mechanism when facing the external interference.In respect of the issues above,the paper carries out in-depth research,and its research contents are shown as follows:(1)A modified force density form-finding method has been proposed to solve the form-finding problem of tensegrity.The characteristics of form-finding problem of tensegrity structure are analyzed,the form-finding problem is transformed into nonlinear unconstrained optimization problem,and solved by conjugate gradient method.Numerical simulations show that the modified force density form-finding algorithm has the advantages of fast calculation speed and high accuracy,which lays a foundation for further investigate and discussion of the form-finding algorithm under the noise polluted.(2)In order to solve the form-finding problem of tensegrity structures under the noise environment,a modified Broyden-Fletcher-Goldfarb-Shanno(MBFGS)algorithm is presented to construct the form-finding algorithm for which could suppress the noise.From the viewpoint of nonlinear unconstrained optimization problem,an approximate Hessian matrix is established as the modified stiffness matrix by MBFGS algorithm to ensure the positive definite of the tensegrity structure during the form-finding process.(3)In view of the development of bionic knee exoskeleton structure,the characteristics of physiological structures such as skeletal,muscle and ligament of human knee joint are analyzed for the purpose of simplifying the knee joint structure.The self-locking function of human knee joint is realized by designing the mechanical structure and the prototype of bionic knee joint structure principle is established.The kinematics of bionic knee joint structure such as singular configuration and working curve are analyzed and the dynamics model of bionic knee joint is designed to lay the foundation for the design of control algorithm.(4)Aiming at the control problems of bionic knee joint,a noise tolerant zeroing neural network(NTZNN)is designed as a controller.The effects of knee joint torque at different rehabilitation stages and fixed noise have been considered in the model,numerical simulations and platform experiments results show that the NTZNN controller has advantages in noise suppression.