Multi-Camera Visual Servoing Formulation to Control a Composite Cartesian-Delta Surgical Robotic Device

During the past two decades, the utilization of vision systems in neurosurgical procedures has become more important. These systems are crucial in minimizing exposure to inoperative radiological imaging and serve as a valuable tool for identifying and addressing injuries. In addition, virtualized instrumented surgical robots can assess the effectiveness of novel surgical approaches compared to conventional and commercial systems. Our primary objective in this study is to estimate the exact pose of the end effector during neuronavigation. To ensure safe neuronavigation, a multi-camera vision system provides real-time feedback for a composite robotic device, the Cartesian-Delta robot, during simulated neurosurgical procedures. Moreover, our visual servoing scheme is integrated as part of an adaptive control law within a virtual environment to estimate the pose of the surgical tool. This design effectively tracks the positions of colored spherical "landmarks" after an initial calibration, offering a significant advantage as this calibration is a one-time process independent of the specific surgical procedure. The numerical results confirm that the implementation of such a visual system can serve as excellent feedback for supervised surgical robots, particularly in scenarios where precise surgical tasks need to be performed within confined spaces, such as trepanation. Compared to traditional approaches, our research’s primary contributions lie in integrating the vision system into the robot control system, the one-time calibration process, and the potential reduction in intraoperative radiological exposure.