A Novel Method for Knot-Tying in Autonomous Robotic assisted Surgery Using Deep Learning

Abstract

With the advent of robot-assisted surgery, surgical task automation, which means that robotic assistants could autonomously execute certain commonly occurring tasks, is more and more appealing and has been studied over the last several years because of the booming of deep learning. Mainly, such partial automation can help reduce the surgeon's workload and allow surgeons to focus more on critical elements of the surgical workflow. In this paper, we propose a novel method using deep learning based on Variational Autoencoders for robotic assistants to learn knot-tying with the Data Set of manual operation, instead of learning from video demonstrations. Taking the circle action of knot-tying as an example, we make use of the VAE network to conduct feature learning and autonomous generation of knot-tying trajectories. During this time, the appropriate VAE network is built and implemented training, and after 100 rounds of training, experimental results show that we successfully acquire trajectories as expected using smaller Data Set with VAE network.