Error Identification and Compensation of a Dual-Robot Measuring and Machining System With an Integrated Visual Sensor

Abstract

To address the growing demand for intelligent manufacturing in personalized, large-scale, and multivariety production, multirobot systems have demonstrated significant potential, as they can tackle complex tasks unattainable for a single robot. A critical challenge for enabling accurate cooperative operations is the precise calibration of relative poses and kinematic parameters of the robot system. In this study, building upon the preliminary calibration results of a dual-robot system, we introduce an error identification and compensation method hinged on data from an integrated visual sensor. Under the premise that the hand-eye, base-base, and tool-end relationships in the dual-robot system have been precalibrated, a novel kinematic error transfer model for the dual-robot system is established to identify and compensate for errors in robot kinematic parameters and the hand-eye, base-base, and tool-end relationships simultaneously, using the same input data as the preliminary calibration. As a result, the cooperative operation accuracy of the dual-robot system can be further enhanced. To demonstrate the feasibility and superiority of the proposed method, four other methods are compared to the proposed method through both simulations and experiments. The comparison results confirm the superiority of the proposed method in terms of accuracy and practicality.