Deep Reinforcement Learning Based Semi-Autonomous Control for Robotic Surgery

EasyChair Preprint 8180

Abstract

Autonomy holds substantial potential in bringing benefits to the robotic surgery industry. In this project, we focused on leveraging deep reinforcement learning methods to automate reaching and grasping operations during a robotic-assisted surgery with a developed handheld controller for manual control. The user study showed that the proposed framework can reduce the completion time (T) by 19.1% and the travel length (M) by 58.7%.

Keyphrases: Deep Reinforcement Learning, robotic surgery, semi-autonomous

@booklet{EasyChair:8180,
  author    = {Ruiqi Zhu and Dandan Zhang and Benny Lo},
  title     = {Deep Reinforcement Learning Based Semi-Autonomous Control for Robotic Surgery},
  howpublished = {EasyChair Preprint 8180},
  year      = {EasyChair, 2022}}