Detail View

Occlusion-Free Optical Sensing Device for Safe Human-Robot Collaboration

Citations

WEB OF SCIENCE

Citations

SCOPUS

Metadata Downloads

DC Field Value Language
dc.contributor.advisor 윤동원 -
dc.contributor.author Yeon Kang -
dc.date.accessioned 2026-01-23T10:59:07Z -
dc.date.available 2026-01-23T10:59:07Z -
dc.date.issued 2025 -
dc.identifier.uri https://scholar.dgist.ac.kr/handle/20.500.11750/59770 -
dc.identifier.uri http://dgist.dcollection.net/common/orgView/200000892321 -
dc.description Human-Robot Collaboration, Occlusion-Free Optical Sensing, Manipulator-Integrated Rotational Sensor, Slip-Ring Mechanism, Time-of-Flight (ToF) Sensing -
dc.description.abstract In this thesis, a rotational optical sensing device that can be embedded or externally mounted on the manipulator link is proposed. The proposed device combines a newly designed slip ring-based mechanism with Time-of- Flight (ToF) sensors, generalizing hardware and software design parameters, and developing an algorithm to convert the detected obstacle positions into a 3D coordinate system. Through the implementation of a vision and reinforcement learning-based human-robot shortest distance measurement algorithm, this study identified that the traditional camera-based vision sensors and deep learning-based position estimation methods have limitations in detection during occlusion situations caused by overlapping robots and humans or obstacles. In contrast, the device proposed in this paper was demonstrated to detect humans and obstacles omnidirectionally without shadowed areas, even with fewer sensors compared to conventional skin-type and ring-type sensors. Finally, the effectiveness of the proposed device and algorithm was experimentally verified through obstacle and human detection experiments in various overlapping scenarios, integrating the vision and reinforcement learning-based algorithm with the proposed RotOptiRing system.|본 논문에서는 매니퓰레이터 링크에 내장 또는 외장 하여 활용할 수 있는 회전식 광학 센싱 장치를 제안하였다. 제안된 장치는 새롭게 고안된 슬립 링 기반 메커니즘과 Time-of-Flight(ToF) 센서를 결합하여, 하드웨어 및 소프트웨어 설계 변수들을 일반화하였으며, 감지된 장애물의 위치를 3차원 좌표계로 변환하는 알고리즘을 개발하였다.
직접 구현한 비전 및 강화학습 기반의 사람-로봇 간 최단거리 측정 알고리즘을 통해, 기존의 카메라 기반 비전 센서 및 딥러닝 기반 위치 추정 방식은 로봇과 사람의 중첩이나 장애물에 의한 차폐(occlusion) 상황에서 탐지에 한계가 있음을 규명하였다. 반면, 본 논문에서 제안한 장치는 기존의 스킨 타입 및 링 타입 센서에 비해 적은 수의 센서만으로도 음영 지역 없이 전방위적으로 사람 및 장애물을 감지할 수 있음을 입증하였다.
마지막으로, 비전 및 강화학습 기반 알고리즘과 제안하는 RotOptiRing을 통합한 시스템에서 다양한 중첩 시나리오에 대한 장애물 및 사람 감지 실험을 통해 제안한 장치와 알고리즘의 유효성을 실험적으로 검증하였다.
-
dc.description.tableofcontents Ⅰ. Introduction 1
ⅠⅠ. Vision-Based Distance Estimation for Human-Robot Collaboration 9
2.1 Vision-Based Instance Segmentation Framework 9
2.1.1 Overview of Vision-Based Approaches 9
2.1.2 Custom Dataset Construction 11
2.1.3 Segmentation Algorithm Evaluation 14
2.2 RGB-D Based Human-to-Robot Distance Estimation 18
2.2.1 Implementation with RealSense D435i in ROS2 Environment 18
2.2.2 Minimum Distance Estimation using Segmentation Mask 20
2.2.3 Limitations in Occlusion Scenarios 25
ⅠⅠⅠ. Ring-Type Optical Sensor for Occlusion-Free Manipulator Control 27
3.1 Overview of the Proposed System 27
3.2 High-Level Collision Avoidance Controller 29
3.2.1 Kinematic Structure of the Manipulator 29
3.2.2 Obstacle Detection Using Ring-Type Sensor 33
3.3 Low-Level Collision Avoidance Controller 35
3.3.1 Safety Zone Definition Based on ISO 15066 35
3.3.2 Velocity and Joint Control under Safety Constraints 38
3.3.3 3D Artificial Potential Field Modeling 40
3.4 Simulation and Experimental Evaluation 44
3.4.1 Simulation in Obstacle-Free and Obstacle-Present Environments 44
3.4.2 Experiment in Obstacle-Free and Obstacle-Present Environments 48
3.4.3 Comparison Between Simulation and Experiment 52
3.4.4 Velocity Comparison with Conventional Method 53
3.5 Chapter 3: Conclusion 55
ⅠV. RotOptiLink: Embedded Rotary Optical Sensor in Manipulator Link 56
4.1 Overview of the Proposed System 56
4.2 Generalization of Design Parameters 58
4.3 Design and Fabrication 63
4.3.1 Mechanical Design of RotOptiLink 63
4.3.2 Fabrication of RotOptiLink 66
4.4 Experiments 71
4.4.1 Validate Design parameters 71
4.4.2 Sensing Performance Test 75
4.4.3 Full-Range Detection with a 3-Axis Manipulator 77
4.4.4 Moving Obstacle Detection Test 80
4.4.5 Comparison with Conventional Sensor Types 83
4.5 Chapter 4: Conclusion 87
V. RotOptiRing: Externally Mounted Rotary Optical Sensor on Manipulator Link 88
5.1 Overview of the Proposed System 88
5.2 Generalization of Design Parameters 91
5.3 Design and Fabrication 99
5.3.1 Mechanical Design of RotOptiRing 99
5.3.2 Fabrication of RotOptiRing 103
5.4 Coordinate Transformation and Visualization 116
5.5 Experiments 122
5.5.1 Validate Design parameters 122
5.5.2 Sensing Performance Test 125
5.5.3 Measurement Accuracy with Obstacle Inclination 127
5.5.4 Comparative Occlusion Test for Human Detection 138
5.5.5 Comparison with Conventional Sensor Types 144
5.6 Chapter 5: Conclusion 147
VI. Conclusion 148
References 150
요약문 155
-
dc.format.extent 155 -
dc.language eng -
dc.publisher DGIST -
dc.title Occlusion-Free Optical Sensing Device for Safe Human-Robot Collaboration -
dc.title.alternative 안전한 인간-로봇 협업을 위한 비전 기반 객체 인식을 보완하는 차폐 회피형 광학 센싱 장치 -
dc.type Thesis -
dc.identifier.doi 10.22677/THESIS.200000892321 -
dc.description.degree Doctor -
dc.contributor.department Department of Robotics and Mechatronics Engineering -
dc.contributor.coadvisor Minho Hwang -
dc.date.awarded 2025-08-01 -
dc.publisher.location Daegu -
dc.description.database dCollection -
dc.citation XT.RD 강64 202508 -
dc.date.accepted 2025-07-21 -
dc.contributor.alternativeDepartment 로봇및기계전자공학과 -
dc.subject.keyword Human-Robot Collaboration, Occlusion-Free Optical Sensing, Manipulator-Integrated Rotational Sensor, Slip-Ring Mechanism, Time-of-Flight (ToF) Sensing -
dc.contributor.affiliatedAuthor Yeon Kang -
dc.contributor.affiliatedAuthor Dongwon Yun -
dc.contributor.affiliatedAuthor Minho Hwang -
dc.contributor.alternativeName 강연 -
dc.contributor.alternativeName Dongwon Yun -
dc.contributor.alternativeName 황민호 -
dc.rights.embargoReleaseDate 2028-08-31 -
Show Simple Item Record

File Downloads

  • There are no files associated with this item.

공유

qrcode
공유하기

Total Views & Downloads

???jsp.display-item.statistics.view???: , ???jsp.display-item.statistics.download???: