목차

Acknowledgements i
Contents ii
List of Figures vi
List of Tables xii
List of Algorithms xiii
List of Abbreviations xiv
Abstract xv
Chapter 1. Introduction 1
1.1 Background 1
1.2 Organization of thesis 5
Chapter 2. Review of formation control systems using omni-directional vision SLAM 8
2.1 Overview 8
2.2 Existing omni-directional video approaches 9
2.2.1 Multi-view video 10
2.2.2 Omni-directional video using parabolic mirror 11
2.2.3 Omni-directional video using fisheye lens 14
2.3 Existing omni-directional vision SLAM approaches 15
2.3.1 Ceiling vision SLAM using general camera 16
2.3.2 Ceiling vision SLAM using fisheye camera 18
2.3.3 Omni-directional SLAM using multi-view camera 19
2.3.4 Global position recognition using omni-directional image 20
2.3.5 Omni-directional SLAM using parabolic mirror 21
2.3.6 Position recognition based on omni-directional SLAM using parabolic mirror 23
2.3.7 Depth extraction using omni-directional stereo vision 24
2.3.8 Global position recognition using omni-directional stereo vision based on single camera 26
2.3.9 Comparison of the existing methods 27
2.4 Existing formation control approaches 29
2.4.1 Behavior-based approach 29
2.4.2 Virtual structure approach 32
2.4.3 Leader-follower approach 33
2.5 Motivation and proposition of novel ideas 34
Chapter 3. Formation Control Systems Using Omni-directional Vision SLAM 37
3.1 Formation Control System Model 37
3.2 The Robot Modeling 38
3.3 The Robot Design 40
3.3.1 The Robot Component 40
3.3.2 Fisheye Camera 43
3.3.3 Controller 44
Chapter 4. Formation Control Algorithm Using Omni-directional Vision SLAM 46
4.1 Overview 46
4.2 Omni-directional vision SLAM based on fisheye camera 49
4.2.1 Removal of the walls and floor areas using the HSI color space 49
4.2.2 Clustering using the labeling 52
4.2.3 Extraction of feature points using LKOF method 56
4.2.4 Correction of locations based on semispherical modeling 59
4.2.5 Local mapping using corrected points of an object 63
4.3 Behavior-based Formation control using global map 65
4.3.1 Estimation of an object characteristic using analysis of vector information 65
4.3.2 Collision avoidance based on curve motion 67
4.3.3 Formation control according to situation recognition 70
Chapter 5. Experimental Results 74
5.1 Global position measure system for experiment 75
5.2 Experiment: Position recognition in static robot environment 80
5.2.1 Experimental environment 80
5.2.2 Experimental result 81
5.3 Experiment: Position recognition in dynamic robot environment 86
5.3.1 Experimental environment 86
5.3.2 Experimental result 87
5.4 Experiment: Avoidance of dynamic obstacle in the rear 91
5.4.1 Experimental environment 91
5.4.2 Experimental result 92
5.5 Experiment: Formation control in static obstacle 96
5.5.1 Experimental environment 96
5.5.2 Experimental result 97
5.6 Experiment: Formation control in dynamic obstacle 101
5.6.1 Experimental environment 101
5.6.2 Experimental result 102
5.7 Experiment: Formation control in wall 106
5.7.1 Experimental environment 106
5.7.2 Experimental result 107
Chapter 6. Conclusion and Future Research 111
References 115
Abstract in Korean 125