A computationally efficient and robust kinematic calibration model for industrial robots with kinematic parallelogram

Abstract

We propose a computationally efficient and robust DH kinematic calibration model to solve the kinematic calibration of the robots with kinematic parallelogram. This model segments the closed chain of kinematic parallelogram into two open chains. Then, by constraining positions of two endpoints in every chain, this model reconstructs kinematic parallelogram as a serial structure. This reconstruction-based DH model (called RST-based kinematic calibration model) has not only proven to be robust and effective in calibration but it is also favored from a computational efficiency viewpoint since it consists of comparatively fewer error parameters. To calculate the error parameters, the genetic algorithm is applied here. The positioning errors on the end-effector (EE) of a Comau Smart5 NJ 220-2.7 robot based on the proposed RST-based kinematic calibration model, modified kinematic calibration model and continuous kinematic calibration model are performed. By comparing the means, standard deviations and computational time of the positioning errors, computational efficiency and robustness of these three models are compared. Finally, the experimental results validate among the three DH calibration models, the proposed RST-based model has the advantages of computational efficiency and robustness. And they successfully validate the feasibility of the present model.