Autonomous high precision positioning of surgical instruments in robot-assisted minimally invasive surgery under visual guidance

Despite the fact that minimally invasive robotic surgery provides many advantages for patients, complex tasks are still time-consuming, error-prone and lead to quicker fatigue of the surgeon. Automating recurrent tasks could greatly reduce total surgery time for patients. While surgeons gain most of the information which is necessary to perform the operation from the visual feedback of cameras, there is only little work on autonomous systems utilizing visual information to generate movement commands. A major step towards automated tasks is the autonomous positioning of surgical instruments with high precision inside the situs. In this paper, we tackle the challenges arising from automated positioning by employing visual servoing techniques in two ways: On one hand, a calibration of all system components has to be performed to enable position-based servoing in Cartesian space. On the other hand, lever effects which appear due to the nature of laparoscopic surgery and intrinsic system imprecisions may not be overcome with calibrations. Therefore, the instruments can also be servoed image-based. Combining both approaches to a switching scheme allows for autonomous high precision positing of surgical instruments in a complex setup with four robots.