Geometric Calibration of Cobots using Circle-Point Analysis for Surgical Applications

The accuracy of the cobot and its frequent monitoring are paramount in robot-assisted surgeries (RAS). Further, on-site assurance without additional infrastructure is preferred. Hence, an in-house method that uses a thermographic camera (commonly used for navigation in RAS) for cobot calibration has been developed. Also, it is hypothesized that the overall system accuracy would be improved as the current work focuses on a good agreement between the camera and the cobot rather than just enhancing the accuracy of the cobot or camera, as in the prior works. With the well-known circle point analysis as the basis, a novel geometric calibration approach with improved identification algorithms was devised. The additional transformations required to accommodate the measurements and validations using ground truths, namely, the cobot's built-in controller and camera, were established. The efficacy of the developed approach was verified on both physical and virtual cobots. The errors in estimating the tool's position and orientation for specified joint displacements were shown to be significantly less with the calibrated parametric values than the nominal ones when using a camera. They were also shown to be on par with studies based on the cobot's built-in controller. Therefore, the proposed approach is promising and can be readily employed for cobot calibration in a surgical environment.