Publications

Detailed Information

Stable Rolling Locomotion for Variable Topology Truss Robot : 가변 토폴로지 트러스 로봇의 안정적인 주행 알고리즘 개발

DC Field Value Language
dc.contributor.advisor김종원-
dc.contributor.author박수민-
dc.date.accessioned2020-10-13T02:38:04Z-
dc.date.available2020-10-13T02:38:04Z-
dc.date.issued2020-
dc.identifier.other000000163254-
dc.identifier.urihttps://hdl.handle.net/10371/169124-
dc.identifier.urihttp://dcollection.snu.ac.kr/common/orgView/000000163254ko_KR
dc.description학위논문 (박사) -- 서울대학교 대학원 : 공과대학 기계공학과, 2020. 8. 김종원.-
dc.description.abstractVariable Topology Truss (VTT) is truss structured modular robot that can self-reconfigure its topology and geometric configuration, which can be usefully applied to rescuing work in disaster site. In this thesis, design of VTT is introduced and stable rolling locomotion algorithm for VTT is proposed.
To achieve self-reconfiguration feature, VTT are composed specially designed members and nodes. VTTs members consist of Spiral Zippers which are novel linear actuators that has high extension ratio, light weight and high strength. VTTs nodes consist of Passive Member-Ends and Master Member-Ends. Passive Member-Ends are linkage type spherical joint with large angle range that can accommodate many members. Master Member-Ends are spherical manipulators that built in Sphere and it move member to change topology of VTT.
Rolling locomotion of VTT is achieved by controlling the center of mass by geometric reconfiguration. However, the locomotion planning is complex problem, because VTT is parallel mechanism with high degree of freedom and many constraints, which makes it difficult to predict and avoid constraints for feasible planning. Thus, it needs stable algorithm that can find locomotion trajectory even in complicated and large environment. In addition, since VTT has many sophisticated components, the algorithm must prevent VTT being damaged from ground by tumbling.
To meet the requirements, proposed locomotion algorithm is composed of 3 steps; support polygon planning, center of mass planning and node position planning. In support polygon planning, support polygon path is planned by newly proposed random search algorithm, Polygon-Based Random Tree (PRT). In center of mass planning, trajectory of desired projected center of mass is planned by maximizing stability feature. Planned support polygon path and center of mass trajectory guide VTT to have good-conditioned shape which configuration is far from constraints and makes locomotion planning success even in complex and large environment.
In node position planning, Non-Impact Rolling locomotion algorithm was developed to plan position of VTTs nodes that prevent damage from the ground while following planned support polygon path and center of mass trajectory.
The algorithm was verified by two case study. In case study 1, locomotion planning and simulation was performed considering actual constraints of VTT. To avoid collision between VTT and obstacle, safety space was defined and considered in support polygon planning. The result shows that VTT successfully reaches the goal while avoiding obstacles and satisfying constraints.
In case study 2, locomotion planning and simulation was performed in the environment having wide space and narrow passage. Nominal length of VTT was set to be large in wide space to move efficiently, and set to be small in narrow passage to pass through it. The result shows that VTT successfully reaches the goal while changing its nominal length in different terrain.
-
dc.description.abstract가변 토폴로지 트러스 (Variable Topology Truss, VTT)는 토폴로지와 기하학적 형상의 재구성이 가능한 트러스 구조의 모듈 로봇이다. 본 논문에서는 VTT의 설계 구조를 소개하고 VTT의 안정적인 주행을 알고리즘을 제안한다.
VTT는 토폴로지와 기하학적 형상의 재구성을 위해 특수한 구조의 멤버와 노드를 가진다. VTT의 멤버는 높은 압축비, 가벼운 중량, 높은 강도를 가진 신개념 선형 구동기인 스파이럴 지퍼로 구성되어 있다. VTT의 노드는 패시브 멤버 엔드와 마스터 엔드로 구성되어 있다. 패시브 멤버는 링키지 구조의 3 자유도 관절로, 넓은 각도 구동 범위를 가지고 있고 많은 수의 멤버를 연결할 수 있다. 마스터 멤버 엔드는 노드 부의 내장된 구형 매니퓰레이터로, 토폴로지 재구성 시 멤버를 이동시키는데 사용된다.
VTT는 기하학적 형상을 변화하여 구르는 움직임을 통해 주행한다. VTT의 주행 알고리즘은 서포트 폴리곤 계획 단계, 무게 중심 계획 단계, 노드 위치 계획 단계로 이루어진다. 서포트 폴리곤 계획 단계에서는 새롭게 제안된 무작위 탐색 (random search) 알고리즘인 Polygon-Based Random Tree (PRT)을 적용해 서포트 폴리곤의 경로를 계획한다. 무게 중심 계획 단계에서는 안정성을 최대화하는 VTT의 무게 중심 궤적을 계획한다. 계획된 서포트 폴리곤 경로와 무게 중심 궤적을 VTT가 제한 조건으로부터 먼 좋은 상태의 형상을 유지하게 하여 복잡한 환경에 대해서도 경로 계획이 실패하지 않고 안정적으로 이루어질 수 있도록 한다.
노드 위치 계획 단계에서는 서포트 폴리곤 경로와 노드 위치의 궤적을 추종하는 노드 위치 궤적을 계획한다. 이 과정에서 비충격 롤링 이동 알고리즘 (Non-Impact Rolling locomotion algorithm)을 적용하여 지면과의 충돌로 인한 충격이 일어나지 않는 궤적을 계획한다.
실제 VTT의 제한 조건을 반영한 모델에 본 알고리즘을 적용하여 시뮬레이션을 수행한 결과, VTT가 모든 제한 조건을 만족하고 장애물을 회피하면서 목표 지점에 도달할 수 있음을 확인하였다.
-
dc.description.tableofcontentsChapter 1. Introduction 1
1.1 Motivation 1
1.2 Previous Truss Type Modular Robot 4
1.3 Previous Research on VTTs Locomotion 8
1.3.1 Heuristic Based Methods 9
1.3.2 Optimization Based Method 10
1.4 Objectives of Locomotion Algorithm 12
1.5 Contribution of Thesis 13
1.5.1 Introduction to Hardware Design of VTT 13
1.5.2 Stable Rolling Locomotion of VTT 15

Chapter 2. Design of Variable Topology Truss 17
2.1 Member Design 18
2.1.1 Spiral Zipper 20
2.1.2 Tensioner 26
2.2 Node Design 28
2.2.1 Passive Member-End and Sphere 29
2.2.2 Master Member-End 36
2.3 Control System 40
2.4 Node Position Control Experiment 44

Chapter 3. Mathematical Model of Variable Topology Truss 47
3.1 Configuration and Terminology 47
3.2 Inverse Kinematics 50
3.3 Constraints 51
3.4 Stability Criteria 64

Chapter 4. Locomotion Algorithm 66
4.1 Concept of Locomotion Algorithm 67
4.1.1 Method for Successful Planning and Obstacle Avoidance 67
4.1.2 Method to Prevent Damage from the Ground 71
4.1.3 Step of Locomotion Algorithm 72
4.2 Support Polygon Planning 73
4.2.1 Polygon-Based Random Tree (PRT) Algorithm 73
4.2.2 Probabilistic Completeness of PRT Algorithm 79
4.3 Center of Mass Planning 85
4.4 Node Position Planning 86
4.4.1 Concept of Non-Impact Rolling Locomotion 86
4.4.2 Planning Algorithm for Non-Impact Rolling Locomotion 89
4.4.3 Optimization Problem of Moving Phase 94
4.4.4 Optimization Problem of Landing Phase 98
4.4.5 Optimization Problem of Transient Phase 99

Chapter 5. Experimental Verification 100
5.1 Case Study 1: Actual VTT Prototype 101
5.1.1 Simulation Condition 101
5.1.2 Obstacle Avoidance Method 103
5.1.3 Simulation Result 104
5.2 Case Study 2: Environment with Narrow Passage 111
5.2.1 Simulation Condition 111
5.2.2 Support Polygon Planning with Varying Nominal Length 114
5.2.3 Simulation Result 117

Chapter 6. Conclusion 126
Bibliography 129
Abstract in Korean 134
-
dc.language.isoeng-
dc.publisher서울대학교 대학원-
dc.subjectvariable topology truss-
dc.subjectvariable geometry truss-
dc.subjectlocomotion planning-
dc.subjecttrajectory optimization-
dc.subjectrandom search algorithm-
dc.subjectobstacle avoidance-
dc.subject모듈 로봇-
dc.subject가변 토폴로지 트러스-
dc.subject주행 경로 계획-
dc.subject장애물 회피-
dc.subject무작위 탐색 알고리즘-
dc.subject.ddc621-
dc.titleStable Rolling Locomotion for Variable Topology Truss Robot-
dc.title.alternative가변 토폴로지 트러스 로봇의 안정적인 주행 알고리즘 개발-
dc.typeThesis-
dc.typeDissertation-
dc.contributor.AlternativeAuthorSumin Park-
dc.contributor.department공과대학 기계공학과-
dc.description.degreeDoctor-
dc.date.awarded2020-08-
dc.identifier.uciI804:11032-000000163254-
dc.identifier.holdings000000000043▲000000000048▲000000163254▲-
Appears in Collections:
Files in This Item:

Altmetrics

Item View & Download Count

  • mendeley

Items in S-Space are protected by copyright, with all rights reserved, unless otherwise indicated.

Share