Synergistic Robotic Surgery: Novel Techniques to Harmonize Surgeon and Robot

During the last 150 years, surgery has evolved from a last resort solution to the highly specialized discipline that it is today. Technological advancement has substantially contributed to this evolution. Thanks to the combined efforts of doctors and engineers, minimally invasive surgery has become a reality, where trauma to the patient is kept so small that recovery is much faster. Yet, this is not the end. The continuing technological progress provides the surgeon with more and more options to improve the quality of treatment. Today, robotic surgery is a prominent representative of this technological innovation. Robots offer the possibility to operate with higher accuracy, speed up the learning process, enhance surgeon comfort and control the motion of complex tools. However, the clinical deployment of these advantages is not straightforward and also substantially differs between surgical disciplines. The focus of this dissertation lies on robotic laparoscopy. Within this medical specialization, telemanipulation systems are presently the standard. With this type of system, the robot copies the movements that the remotely located surgeon executes with a sophisticated joystick. So far, it has however proven difficult to justify the hefty price tag of these robots. The existing systems only employ a fraction of the full potential of robotics. Especially the opportunity to embed intelligent behavior in these robots is hardly exploited. At the same time, the surgeon needs to make compromises to integrate the system by completely revisiting his/her normal approach and workflow. This can for instance be witnessed by the long set-up time of the devices and the distance they create between the surgeon, the team and the patient. This dissertation aims to reduce the distance, both physically and psychologically, that exists between surgeon and robot. It proposes methods that optimize the synergy between both. This allows the surgeon to operate more effectively and safely, supported by a robot that provides assistance in a way that seamlessly fits in with his/her expectations. Based on this starting point, a number of important design guidelines for synergistic robotic surgery are formulated. They promote the comanipulation principle for robotic laparoscopy. In comanipulation, the surgeon and the robot jointly hold and manipulate the surgical instruments, in contrast to the remote control in telemanipulation. To further reduce the impact of the comanipulation system on the surgical practices, it is of importance that the robot can be moved without resistance and that it can automatically align itself with the incision point in the patient. If synergy is to be guaranteed, a careful approach is also required towards robot perception, which is necessary for more advanced robot assistance. The surgeon cannot be burdened with extra steps to collect or interpret data. While these design guidelines are valuable for the surgeon, their practical implementation implies a number of technical challenges, which are studied and tackled in the remainder of this dissertation. The outcomes are every time applied to the particularly demanding discipline of fetal surgery. The first part of the challenges is related to the incision point in laparoscopy and has an impact on the robot control strategy. To ensure that the robot can automatically align itself with the incision point, a robust and responsive method is developed to measure the position of the incision point. Knowledge of this point is subsequently utilized to enhance the performance of the proposed type of comanipulation robot. On the one hand, it is shown how this knowledge of the incision point allows for improved accuracy, precision and bandwidth of instrument tip position measurements. On the other hand, an algorithm is designed that applies this knowledge to expand the stability margins of the robot. The combination of the developed techniques results in safer, more effective and more synergistic robot assistance. The second part of the challenges is related to robot perception. Perception-related problems are elaborated for two clinical use cases: contactless surgery and endoscope guidance. For these use cases, sensors are selected that can be integrated in laparoscopic instruments suitable for fetal surgery. Techniques to process the sensor data into actionable computer models are then developed. These techniques work entirely automatically, without requiring the attention of the surgeon to collect, enrich or verify the data. Through a series of in-vitro and in-vivo experiments, the feasibility of robot assistance for both clinical cases is demonstrated. From this dissertation, it thus follows that intelligent and effective robot assistance can be successfully realized while taking into account all design guidelines for synergy, even for demanding fetal interventions.status: publishe