| In recent years,flexible robots have attracted wide attention due to their high degree of freedom,adaptability,and high force-to-weight ratio.In the field of engineering applications,flexible robots can change their shapes to suit the environment.In interaction with humans,flexible robots are more safe,because their own elasticity can greatly reduce the concentration stress of collisions and reduce potential safety hazards.However,there are still many difficulties in the study of flexible robots.For example,it is difficult to establish an accurate model,the perception accuracy is not enough,soft robot control is complicated,and so on.In this paper,the static and dynamic models of the humanoid manipulator based on flexible joints were established,and the sensing,driving and controlling methods were given,and verified by simulation and experiments.The main work of this paper includes:Firstly,the thesis established the mechanical model of a humanoid robot.Based on specific assumptions,the dynamics and kinematics model were established.A static model was established,which represents the relationship between the loading and deformation of the joint under large deformation.The method of sensing the shape and force of flexible joints and fingers was given?Secondly,a platform of the humanoid robotic manipulator was completed by dividing into four parts,control system,force-motion sensing,drive and transmission system and host computer software system.Then,the control model of the humanoid robot hand was established.The relationship between inputs and outputs of the parallel driving mechanism was formulated.The principle of stable contact detection based on strain rates was proposed.A flexible finger feedback control model was given,including displacement control and contact force control.Finally,the system tests of humanoid manipulators were completed,including stable contact detection experiment,finger angle control experiment,and finger contact force control experiment. |
PDF Full Text Request