Title

Optimal Handover of Surgical Needle Using Reinforcement Learning for Suturing Automation

DGIST Authors

Hakyoon Lee ; Minho Hwang ; Sang Hyun Park

Advisor

황민호

Co-Advisor(s)

Sang Hyun Park

Issued Date

2024

Awarded Date

2024-02-01

Citation

Hakyoon Lee. (2024). Optimal Handover of Surgical Needle Using Reinforcement Learning for Suturing Automation. doi: 10.22677/THESIS.200000730097

Type

Thesis

Description

Surgery Automation; Suturing Automation; Reinforcement Learning; Deep Q-learning; Autonomous Regrasping; Visual servo control; Robot calibration

Table Of Contents

Ⅰ. Introduction 1
1.1 The Necessity of Suturing Automation 1
1.2 Handover Automation Methods 1
1.3.1 Testing in a Real-World Environment 1
1.3.2 Testing in a Real-World Environment 2
1.3.3 Testing in a Real-World Environment · 2

Ⅱ. Surgical robot system 5
2.1 robot system component 5
2.2 Mechanical design 5
2.3 Remote Center of Motion mechanism 8
2.4 Remote Kinematics · 8

Ⅲ. Calibration 11
3.1 RGBD camera calibration · 11
3.2 Robot position control error calibration · 16

Ⅳ. Reinforcement Learning algorithm 17
4.1 Deep Q-Learning algorithm 17
4.2 State definition 20
4.3 Action definition 21
4.4 Reward definition 21
4.5 Model experiment · 22

Ⅴ. Real robot test 26
5.1.1 Task space control 26
5.2.1 Visual Servo control 29
5.2.2 Needle pose estimation 30
5.2.3 Needle grasping using RGBD camera 32
5.2.4 Experimental result · 32

VI. Conclusion · 34

URI

http://hdl.handle.net/20.500.11750/48072
http://dgist.dcollection.net/common/orgView/200000730097

DOI

10.22677/THESIS.200000730097

Degree

Master

Department

Department of Robotics and Mechatronics Engineering

Publisher

DGIST

Show Full Item Record