Moccia, Rocco (2021) Vision-Based Autonomous Control in Robotic Surgery. [Tesi di dottorato]

Preview

Text Rocco_Moccia_33.pdf Download (35MB) | Preview

Collection description

Robotic Surgery has completely changed surgical procedures. Enhanced dexterity, ergonomics, motion scaling, and tremor filtering, are well-known advantages introduced with respect to classical laparoscopy. In the past decade, robotic plays a fundamental role in Minimally Invasive Surgery (MIS) in which the da Vinci robotic system (Intuitive Surgical Inc., Sunnyvale, CA) is the most widely used system for robot-assisted laparoscopic procedures. Robots also have great potentiality in Microsurgical applications, where human limits are crucial and surgical sub-millimetric gestures could have enormous benefits with motion scaling and tremor compensation. However, surgical robots still lack advanced assistive control methods that could notably support surgeon's activity and perform surgical tasks in autonomy for a high quality of intervention. In this scenario, images are the main feedback the surgeon can use to correctly operate in the surgical site. Therefore, in view of the increasing autonomy in surgical robotics, vision-based techniques play an important role and can arise by extending computer vision algorithms to surgical scenarios. Moreover, many surgical tasks could benefit from the application of advanced control techniques, allowing the surgeon to work under less stressful conditions and performing the surgical procedures with more accuracy and safety. The thesis starts from these topics, providing surgical robots the ability to perform complex tasks helping the surgeon to skillfully manipulate the robotic system to accomplish the above requirements. An increase in safety and a reduction in mental workload is achieved through the introduction of active constraints, that can prevent the surgical tool from crossing a forbidden region and similarly generate constrained motion to guide the surgeon on a specific path, or to accomplish robotic autonomous tasks. This leads to the development of a vision-based method for robot-aided dissection procedure allowing the control algorithm to autonomously adapt to environmental changes during the surgical intervention using stereo images elaboration. Computer vision is exploited to define a surgical tools collision avoidance method that uses Forbidden Region Virtual Fixtures by rendering a repulsive force to the surgeon. Advanced control techniques based on an optimization approach are developed, allowing multiple tasks execution with task definition encoded through Control Barrier Functions (CBFs) and enhancing haptic-guided teleoperation system during suturing procedures. The proposed methods are tested on a different robotic platform involving da Vinci Research Kit robot (dVRK) and a new microsurgical robotic platform. Finally, the integration of new sensors and instruments in surgical robots are considered, including a multi-functional tool for dexterous tissues manipulation and different visual sensing technologies.

Downloads

Actions (login required)

View Item