Title Page
Contents
Abstract 10
Chapter 1. Introduction 13
1.1. Background and purpose of research 13
1.2. Structure of research 23
Chapter 2. Literature review 26
2.1. Previous studies on collision risk 27
2.2. Previous studies on collision risk considering external factors 29
Chapter 3. Methodology 35
3.1. Fuzzy expert system 35
3.1.1. Rule-based fuzzy reasoning system 36
3.1.2. Fuzzy reasoning process 39
3.1.3. Sugeno-style fuzzy inference 48
3.2. Fuzzy C-Means clustering method 50
3.3. Analytical hierarchy process and fuzzy structural modelling 54
3.3.1. AHP method 54
3.3.2. FSM method 55
Chapter 4. Navigational collision risk solving system 57
4.1. Introduction 57
4.2. Data analysis of vulnerability 60
4.2.1. Traffic situations along Korean coastline 61
4.2.2. Static and dynamic analysis of narrow waterways 64
4.2.3. Marine accidents and number of passing vessels in narrow waterways 67
4.2.4. Current speed and congestion degree in narrow waterways 72
4.2.5. Analysis of fishing boat activities 73
4.3. Framework of navigational collision risk solving system 74
4.3.1. Basic collision risk 76
4.3.2. Solutions for the factors of vulnerability 79
4.3.3. Combined vulnerability 94
4.3.4. Navigational collision risk 96
4.4. Application of navigational collision risk solving system 97
4.5. Summary 102
Chapter 5. Conclusions and recommendations 104
5.1. Conclusions 104
5.2. Discussions and future studies 105
References 106
Table 3-1. Reasoning rules of collision risk 43
Table 3-2. Abbreviations for fuzzy rules 43
Table 3-3. Adopted scale for importance 55
Table 4-1. Marine Accidents Analysis 2013-2015 60
Table 4-2. Number of marine accidents by ship type (2008-2016) 61
Table 4-3. Reasoning rules of DCPA, TCPA and collision risk 78
Table 4-4. Reasoning rules of vulnerability for bad weather 81
Table 4-5. The vulnerability (V) results of testing tuning algorithm 82
Table 4-6. Reasoning rules of vulnerability for strong tidal current 84
Table 4-7. Reasoning rules of vulnerability for accidents prone area 86
Table 4-8. Reasoning rules of vulnerability for traffic congestion 88
Table 4-9. Reasoning rules of vulnerability for operator fatigue 90
Table 4-10. Reasoning rules for vulnerability for fishing beats operating area 93
Table 4-11. Importance matrix of vulnerability factors 95
Table 4-12. Random consistency index 95
Table 4-13. Weight and priority of vulnerability factors 96
Table 4-14. Reasoning rules of navigational CR 96
Table 4-15. Details of target ships in the vicinity of own ship 98
Table 4-16. The vulnerability inputs and results of bad weather 99
Table 4-17. The vulnerability inputs and results of strong tidal current 99
Table 4-18. The vulnerability inputs and results of accident prone areas 99
Table 4-19. The vulnerability inputs and results of traffic congestion 99
Table 4-20. The vulnerability inputs and results of operator fatigue 99
Table 4-21. The vulnerability inputs and results of fishing area 100
Table 4-22. The results of navigation vulnerability factors 100
Table 4-23. Results of the collision risk 101
Figure 1-1. Three types of sea state 16
Figure 1-2. Factors selected for integration of e-Navigation 17
Figure 1-3. The framework designed for high risk ships 19
Figure 1-4. Details included in the process 20
Figure 1-5. Awareness system for vulnerability of vessel 21
Figure 1-6. Navigation risk system based on vulnerability of vessel 22
Figure 3-1. DCPA and TCPA between two approaching vessels 38
Figure 3-2. Linguistic input and output variables 40
Figure 3-3. Crisp inputs of DCPA and TCPA 41
Figure 3-4. The AND product fuzzy process 45
Figure 3-5. Aggregation of the rule outputs 46
Figure 3-6. Sugeno's defuzzification method using fuzzy singleton 49
Figure 3-7. Fuzzy clustering method 51
Figure 3-8. FCM step-by-step flow chat 52
Figure 3-9. Hierarchy of AHP method 54
Figure 4-1. Marine casualties (2016-2017) 63
Figure 4-2. The center of cluster case 1 67
Figure 4-3. The center of cluster case 2 68
Figure 4-4. The center of cluster case 3 69
Figure 4-5. The center of cluster case 4 (0am-6am) 70
Figure 4-6. The center of cluster case 4 (6am-12pm) 70
Figure 4-7. The center of cluster case 4 (12pm-6pm) 71
Figure 4-8. The center of cluster case 4 (6pm-0am) 71
Figure 4-9. The center of cluster case 5 73
Figure 4-10. Clustering result between 15:00 and 15:03 on May 5 2018 74
Figure 4-11. Structure of the navigational collision risk solving system 76
Figure 4-12. Membership functions for DCPA, TCPA and basic CR 78
Figure 4-13. Fuzzy logic membership functions for bad weather 81
Figure 4-14. Membership functions for the length (m) of ship 82
Figure 4-15. Tuning values based on the size of ships (length) 83
Figure 4-16. Fuzzy logic membership functions for strong tidal current 84
Figure 4-17. Fuzzy logic membership functions for accidents prone area 86
Figure 4-18. Fuzzy logic membership functions for traffic congestion 88
Figure 4-19. Fuzzy logic membership functions for operator fatigue 90
Figure 4-20. Fuzzy logic membership functions for fishing beats operating area 93
Figure 4-21. Fuzzy membership functions for module four 97
Figure 4-22. Navigational traffic situation 98