Efficient TCP Calibration Method for Vision Guided Robots Based on Inherent Constraints of Target Object

Abstract

Tool Center Point (TCP) calibration and target object calibration are essential to guarantee the accuracy of Vision Guided Robot (VGR) systems. After calibration, the robot can know the object's position and orientation from the vision system and then move the TCP to a target point. However, conventional calibration methods are time-consuming and often resort to external tools. We propose a universal method based on the inherent constraints of the target points and use it to simultaneously calibrate the TCP and the target object. In order to obtain the TCP parameters, first, we build a constraint model to TCP by exploiting the target object calibration. By this means, TCP calibration can be combined into the target object calibration, and hereby no external calibration tools are required. Second, we represent this model as an optimization problem of minimizing the reprojection error in the domain of Lie algebra. Third, we solve the numerical problem by the Gauss-Newton algorithm with the perturbation model. Notably, we point out that the universal model can be reduced to a particular simple case when a specific point on the object is available, e.g., a corner of the target object can be recognized by the camera. Here, we present a particular coordinate conversion method to exempt the case from calculating the TCP parameters, which is applicable in a wide range of applications. The practicability and the accuracy of the proposed methods are verified by comparative experiments and numerical simulations. Results show that our method improves TCP calibration efficiency and accuracy by integrating the TCP calibration and object calibration.