Identifiable Parameter Analysis for the Kinematic Calibration of a Hybrid Robot

In this paper, a statistical method for the determination of the identifiable parameters of a hybrid serial-parallel robot IWR (Intersector Welding Robot) is presented. This method is based on the Markov Chain Monte Carlo (MCMC) algorithm to analyze the posterior distribution and correlation of the error parameters. Differential Evolution algorithm is employed to search a global optimizer as initial values for the random sampling of MCMC. The robot under study has ten degrees of freedom (DOF) and will be used to carry out welding, machining, and remote handing for the assembly of vacuum vessel of the international thermonuclear experimental reactor (ITER). In this paper, a kinematic error model which involves assembling and manufacturing error parameters is developed for the proposed robot. Based on this error model, the mean values of the unknown parameters are statistically analyzed and estimated using the proposed method. Computer simulations reveal that all the reduced independent kinematic parameters can be identified with the complete pose measurements. Results also demonstrate that the identification method is robust and effective with the given measurement noise.