Design And Motion Control Of A Prosthetic Wrist Driven By Dielectric Elastomers

Human wrist is an important link between forearm and palm,which can flexibly adjust the different posture of palm relative to forearm.At present,most prosthetic wrist is based on motor drive,through the connecting rod and other transmission mechanism to transfer power,prosthetic wrist flexibility is relatively poor,and the quality and size of large.In this paper,a new type of dielectric elastomer is used to drive wrist motion to improve the existing problems in wrist research.In this paper,the structure design,kinematics and dynamics modeling,motion control and other aspects of prosthetic wrist driven by dielectric elastomer are systematically studied.Firstly,the physiological structure and motion characteristics of human wrist were studied,and a 3-DOF hybrid 4LS+U prosthetic wrist mechanism was proposed.The parallel mechanism is driven by two pairs of EAP drivers.The rotating mechanism is driven by small motor and connected in series with parallel mechanism.The motion capacity of the prosthetic wrist was analyzed through kinematic simulation,the bearing capacity of the structure was tested by means of finite element analysis,and the structural parameters of the cylindrical actuator were optimized through experiments,and good actuator characteristics were obtained.Secondly,the kinematic models of cartesian space and drive space of the mechanism were established respectively by the improved DH method and vector method,and the position,velocity and acceleration of the wrist mechanism were analyzed.The correctness of the kinematic model was verified by kinematic simulation,and the workspace of the wrist end was solved.Based on the duality relation between Lagrange equation and velocity Jacobian matrix and force Jacobian matrix,the system dynamics model was established and viscous friction was introduced into the drive space.Combined with the electromechanical coupling model of EAP driver,the system dynamics was simulated and analyzed,and the steady-state relationship between wrist Angle and wrist output torque on voltage U was obtained.Then,RBF neural network terminal sliding mode control algorithm is used to design the controller of prosthetic wrist mechanism.The sliding mode surface,control rate,weight updating rate and adaptive compensation rate of RBF neural network are designed and studied,and the stability and convergence of the wrist control system designed are analyzed.The approximation and compensation ability of OLS learning algorithm for neural network sliding mode controller is studied.The important parameters of the learning algorithm are selected and optimized by simulation analysis.Two control groups of controllers were designed to simulate and compare the wrist track tracking performance under the three groups of controllers.The control simulation results show that the RBF terminal sliding mode control algorithm has higher tracking accuracy,faster response speed,smaller steady output chattering,and more prominent control effect.Finally,a prosthetic wrist experiment platform was built to test the movement ability of the prosthetic wrist,and the ultimate Angle of each degree of freedom direction was studied.The dynamic and steady-state characteristics of the rotation Angle and the output plugging moment were experimentally studied for three groups of different electrification modes under no-load and loaded conditions.Combined with RBF neural network terminal sliding mode control algorithm,the trajectory tracking performance of prosthetic wrist terminal was studied,and the effectiveness of the proposed control method and the correctness of the dynamics model were verified.