KidsArm — An image-guided pediatric anastomosis robot

Abstract

Minimally invasive surgery (MIS) revolutionized surgery by drastically reducing patient recovery times by allowing surgeons to perform procedures through a series of small incisions. However, MIS has also increased the complexity of the tasks as tools did not have the same degrees of freedom and dexterity compared to open procedures. In particular, pediatric patients pose a unique challenge as they have smaller volumes and different tissue properties. Our group designed KidsArm, an image-guided pediatric surgical robot, to automate anastomosis. KidsArm is single port anastomosis tool that uses a pair of stereo cameras to generate a 3D point cloud to guide the tool tip and apply a series of sutures. The system was designed to be minimally invasive and constrained by standard pediatric trocar sizes while also being automated. An image processing system was created to extract and track surface features on the simulated tissue samples while providing feedback to the robot controller. The system was tested on two scenarios: side-to-side and end-to-end silicone samples. KidsArm successfully applied 3 sutures autonomously on the side-to-side scenario however the end-to-end scenario proved to be more difficult due to greater deformation and workspace restrictions. However, KidsArm demonstrates that it is feasible for a robot to autonomously perform anastomosis. More work will be required to accelerate the process and characterize the behavior with tissue samples.