Research On Inverse-kinematics And Coordinated Motion Control For Humanoid Multi-fingered Hand

Based on multi-dimensional fingertip force sensing and synchronization control of humanoid multi-fingered hand(Grant No.51175106),the dissertation studies the inverse-kinematics and control of cooperative motion for humanoid multi-fingered hand.The overall performance of multi-fingered hand depends on coordinating property of whole fingers and whole joints of fingers.Therefore,if multi-fingered hand has poor coordinating property,the overall performance of system will be decreased.In order to solve the problem above,the dissertation first proposes the inverse-kinematics algorithm of finger with coupled joints,then studies the synchronization error among all joints of a finger and the synchronous control method among all fingers.Next,the dissertation adds cooperative motion laws of human hands to controller design so that multi-fingered hand is able to work in coordination.Firstly,in allusion to finger with coupled joints,the dissertation proposes a power series based inverse-kinematics algorithm,by which the nonlinear equation including trigonometric transcendental function can be converted into an algebraic equation.The algorithm uses 1-D linear interpolation to compensate errors,which reduces the requirement of transmission ratio accuracy for nolinearly coupled joints.Secondly,through HIT/DLR Hand II,the dissertation analyzes the impacts on position control from different trajectory plannings.A kalman filter based adaptive friction compensation method is added to synchronization error among all joints of a finger.Finally,the dissertation proposes a synchronous control method based on cooperative motion laws of human hands,in which the coupling ratios of joints based on joint coupling ratios of human hands is added to planning layer.And the synchronization error in consideration of synchronism among all joints of a finger and all fingers is designed in control layer,which also includes cooperative motion laws of human hands.At last,according to the results of impedance control experiments,the multi-fingered hand can reproduct grasping motions of human hands availably to accomplish collaborative motion.