An Accuracy Assessment of Hand-Eye Calibration Techniques in Uncertain Environments for Vision Guided Robots

The hand-eye calibration accuracy of a vision-guided robot plays a huge role in the decision-making process of the robot during operation. The majority of work focuses on passive calibration where the calibration is done with a batch of data prior to putting the robot in service. In a previous study, "Accuracy evaluation of hand-eye calibration techniques for vision-guided robots, (2022)", the accuracy of six commonly used passive calibration methods was investigated and showed good performance in terms of small rotation and translation errors. However, that investigation was carried out assuming no uncertainties in the robot operating environment. In this study, we extend that investigation by evaluating the accuracy of those six methods when the robot is operating in uncertain environments. By introducing changes in calibration parameters to mimic uncertain environments, the accuracy of the six methods deteriorates significantly suggesting a need to address this issue. We suggest in our future work, the probable deployment of active calibration to improve the accuracy and discuss some expectations and limitations associated with it.