Research On Peg-in-hole Assembly By Robot Hand-arm System

The hand/arm system gives the robot the ability to perform complex manipulation tasks likes human hand-arm,and using human operation experience directly.Compared to 6-DOF manipulator,the hand-arm system provides better adaptability and operation flexibility.However it also put higher requirements on kinematics analysis,coordinate control and so on.In order to control the hand-arm system,the kinematic model of the robot is established,including the D-H coordinate systems,the forward kinematic model of the 7-DOF manipulator,the inverse kinematic model of the manipulator using “Arm Angle”,and the kinematics models of the dexterous hand.And the joint control strategy of the hand-arm system is also given.A pivotal ability of the hand-arm system is manipulation.The manipulation plan is given in the form of kinematic model which mapping from the target pose of object to the position of joint angles.To improve the precision of manipulation,an online 3-D object reconstruction method is given,which can help the dexterous hand get the pose of the grasped object through a few points information on the surface.To improve the grasp stability during manipulation,a stability index is provided,and using this index contact points is optimized.Besides the workspace of the dexterous hand while grasping a cylinder is analyzed.Peg-in-hole assembly is a typical contact task,which puts a higher requirements on sensing,information analysis and manipulation ability of robot.In this paper,the contact formation graph of round peg and hole is analyzed,and the contact formation graph is augmented with a similarity index in wrench space.A T-S fuzzy classifier is used to identify the contact formation through wrench information,and the parameters of T-S model are computed using the Gravitational Search-based Fuzzy Clustering Algorithm.Finally,a peg-in-hole assembly experiment system are constructed.Artificial data and actual measurement data are used to test the model reconstruction algorithm.And wrench data of each contact formation are measured in experiment to compute the similarity index and test the T-S fuzzy classifier.And at last a peg-in-hole assembly experiment is carried out.