Development of the methods to identify dynamic parameters of the robot arm is necessary for the advanced control and design of the robot systems. The dynamic parameters include the pseudoinertia matrix elements for each link and Coulomb friction and viscous friction coefficients in each joint.
In this thesis, a new method to identify these parameters with rotary joints is proposed. The salient feature of this method is characterized by the adaption of Balancing Mechanism which is used to identify accurately the gravity-related parameters whose identification accuracy affects that of the other parameters significantly. The method consists of four serial experimental procedures from joint 3 to joint 1 without decomposing the robot system into parts after the manufactured Balancing Mechanism is equipped with the unbalanced robot. Data required are joint angles, velocities, accelerations and joint torques and these data are processed by least square identification algorithm to identify the dynamic parameters. The identification method proposed in this thesis is applied to the PUMA-760 robot and we showed that the simulation results are in good agreements with those of the experimental ones which implies that the proposed identification method is powerful and that the identified model parameters are accurate enough to predict the PUMA-760 manipulator motion.
