Isomorphic Design,Human-Like Trajectory Generation,and Adaptive Control Of Upper Limb Exoskeleton

For patients with upper limb motor dysfunction caused by cerebrovascular or Parkinson’s,the patients cannot fully rehabilitation affected by the influences of the uneven personal level of physicians in traditional one-to-one therapy.Thus,the patients cannot independently complete activities of daily living.Upper limb exoskeleton can assist movement by wearing on the outside of upper limb to improve the life quality of the patients.The upper limb exoskeleton is required to naturally assist patients due to the existence of human-exoskeleton physical interaction.Therefore,it is of great significance to develop an isomorphic exoskeleton and research human-like motion.The motion characteristics of the shoulder,elbow,and wrist were analyzed based on anatomy of the human upper limb.A 4-DOFs upper limb exoskeleton was proposed based on human upper limb kinematics and isomorphic design principle.Modular design method was used to design the motion unit of shoulder and elbow,and Bowden-cable driven system.The coordinate system of the exoskeleton was established and the forward kinematics was analyzed based on the D-H parameter method.We analyzed the workspace of exoskeleton by Monte Carlo algorithm to verify the rationality of exoskeleton.The singularity was analyzed by the inverse kinematics,which provided theoretical basis for avoiding the singular position in motion control.We used ADAMS software for dynamic simulation to calculate the maximum driving torque of each joint,which provided a basis for motor selection.The activities of daily living of human upper limb were classified and analyzed.The typical movements of drinking and touching nose were selected to research.We proposed a human-like motion trajectory generation method based on optical motion capture.Firstly,the kinematics model of human upper limb was established based on marker-cluster method,and the position information of the markers was collected by optical motion capture system.Then,the Euler angle calculation method and inner product method were used to calculate the human-like trajectories of the multi-DOFs shoulder joint and single-DOFs elbow joint.Finally,the human-like trajectory was mapped to the exoskeleton joint space from human joint space by a one-to-one mapping method.The control system of upper limb exoskeleton was established and the prototype was developed.The fuzzy adaptive PID control algorithm was designed,and compared with the traditional PID control algorithm for human-like trajectory tracking.The experimental results of trajectory tracking demonstrated that the fuzzy adaptive PID control can reduce the 4 degrees tracking error of traditional PID control to less than 1 degree.To evaluate the human-like similarity when moving by the human-like trajectory.The interaction forces at physical human-exoskeleton connection interfaces were detected when moving by the proposed human-like trajectory or the traditional five-order polynomial trajectory to quantitatively evaluate the human-likeness.The human-like evaluation experimental results indicated that the proposed human-like trajectory generation method can decrease the interaction forces,and provide comfort and natural movement for the wearer.