표제지
목차
Abstract 10
제1장 서론 13
1.1. 연구 배경 및 목적 13
1.2. 연구 방법 및 구성 17
제2장 선박 충돌 예방 지원에 관한 선행 연구 고찰 21
2.1. 선박 충돌 예방 지원 시스템 연구 21
2.2. 선박 충돌 예방 알람 시스템 디자인 연구 35
2.4. 선행 연구 분석 결과에 대한 고찰 39
제3장 Risk Informed Classification Model(RICM) 개발을 위한 연구 방법 40
3.1. Encounter Classification Model(ECM)을 이용한 선박 조우 상황 인식 41
3.2. Logistic Regression 47
3.2.1. 통계적 이론 50
3.2.2. Logistic Regression Coefficient 추정 52
3.3. Logistic Regression을 위한 학습 데이터 54
제4장 Risk Informed Classification Model(RICM) 개발 61
4.1. Encounter Classification Model(ECM) 선택 61
4.2. Logistic Regression 분석 및 RICM 모델 개발 결과 62
4.2.1. Overtaking 조우 상황 62
4.2.2. Crossing 조우 상황 66
4.2.3. Head-on 조우 상황 71
4.2.4. Ambiguous 조우 상황 75
제5장 Risk Informed Classification Model(RICM) 평가 및 분석 80
5.1. Risk Informed Classification Model(RICM) 경향 분석 80
5.2. Risk Informed Classification Model(RICM)에 의한 해상교통 영향 분석 97
5.3. 선박운항자 만족도 평가 104
제6장 결론 106
참고 문헌 108
부록 113
Table 1.1. Cause of collision at the Korea Sea(2015~2019 year Judgment completed) 15
Table 1.2. Ship to ship distance recognized for the first time in a collision... 15
Table 1.3. Ship collision risk information step 17
Table 2.1. Influencing factor of ship situation 21
Table 2.2. Interpretation of the different risk levels in Risk Informed... 22
Table 2.3. Navigational collision risk scale 29
Table 2.4. Risk scores for risk level 30
Table 2.5. Coordinated collision avoidance action index 37
Table 3.1. Encounter classification model angle definition 42
Table 3.2. Classification of the encounter type according to ECM 42
Table 3.3. B₁/AM, B₂, B₃ & H1~H7 ECM figure 1 44
Table 3.4. B4, B5, B6 & H1~H7 ECM figure 2[이미지참조] 45
Table 3.5. B7, B8 & H1~H7 ECM figure 3[이미지참조] 46
Table 3.6. Dynamic & static variable for RICM 56
Table 3.7. Survey scenario design for RICM 60
Table 4.1. The encounter classification for RICM 61
Table 4.2. Non-alarm/Alarm classification model result (Overtaking) 63
Table 4.3. Information/Warning classification model result (Overtaking) 65
Table 4.4. Non-alarm/Alarm classification model result (Crossing) 68
Table 4.5. Information/Warning classification model result (Crossing) 70
Table 4.6. Non-alarm/Alarm classification model result (Head-on) 72
Table 4.7. Information/Warning classification model result (Head-on) 74
Table 4.8. Non-alarm/Alarm classification model result (Ambiguous) 76
Table 4.9. Information/Warning classification model result (Ambiguous) 78
Table 5.1. Overtaking RICM result 1 82
Table 5.2. Overtaking RICM result 2 83
Table 5.3. Crossing RICM result 1 84
Table 5.4. Crossing RICM result 2 85
Table 5.5. Head-on RICM result 1 86
Table 5.6. Head-on RICM result 2 87
Table 5.7. Ambiguous RICM result 1 88
Table 5.8. Ambiguous RICM result 2 89
Table 5.9. Likert scale(5 score) 104
Table 5.10. Ship information & score results 105
Fig. 1.1. Marine Accident at The Korea Sea(2015~2019 year) 14
Fig. 1.2. RICM development method & structure 19
Fig. 2.1. Escalation state concept in MTCAS 23
Fig. 2.2. Principle of overlapping maneuvering... 24
Fig. 2.3. Procedure of a monitoring system of collision... 27
Fig. 2.4. Flow chart of collision prevention algorithm for small fishing vessels 28
Fig. 2.5. Risk scores for subjective risk levels 30
Fig. 2.6. Scenario of ship collision prevention support system 34
Fig. 2.7. Scenario of ship collision prevention support system 34
Fig. 2.8. Three protective layers of elimination of conflict in good visibility 35
Fig. 2.9. The four stages of the encounter process relation to... 36
Fig. 2.10. Coordinated collision avoidance action 37
Fig. 2.11. Stages of the decision making process 38
Fig. 2.12. Graphical user interface of NAVDEC 38
Fig. 3.1. Risk informed classification model(RICM) Step 40
Fig. 3.2. Model for classifying ship encounters according to... 41
Fig. 3.3. Example decision tree(toy data set) 48
Fig. 3.4. Example KNN 48
Fig. 3.5. Example SVM 49
Fig. 3.6. Logistic Regression 51
Fig. 3.7. Non-Alarm, Information, Warning RICM procedure 54
Fig. 3.8. Obtain learning data for Risk Informed Classification Model(RICM) 55
Fig. 3.9. Simple simulation test 56
Fig. 3.10. Simple simulation test result 57
Fig. 3.11. Survey data sheet sample 58
Fig. 4.1. Residual analysis for Non-alarm/Alarm classification (Overtaking) 62
Fig. 4.2. Residual analysis for Information/Warning classification (Overtaking) 64
Fig. 4.3. Residual analysis for Non-alarm/Alarm classification (Crossing) 67
Fig. 4.4. Residual analysis for Information/Warning classification (Crossing) 69
Fig. 4.5. Residual analysis for Non-alarm/Alarm classification (Head-on) 71
Fig. 4.6. Residual analysis for Information/Warning classification (Head-on) 73
Fig. 4.7. Residual analysis for Non-alarm/Alarm classification (Ambiguous) 75
Fig. 4.8. Residual analysis for Information/Warning classification (Ambiguous) 77
Fig. 5.1. Comparison of trends(own ship length) 90
Fig. 5.2. Comparison of trends(target ship length) 91
Fig. 5.3. Comparison of trends(own ship speed) 92
Fig. 5.4. Comparison of trends(target ship speed) 93
Fig. 5.5. Comparison of trends(DCPA) 94
Fig. 5.6. Comparison of trends(TCPA) 95
Fig. 5.7. Comparison of trends(range) 96
Fig. 5.8. Traffic information analysis range(Mok-po area) 97
Fig. 5.9. Elliptical ship domain 98
Fig. 5.10. LTE-M & AIS Ships(minimum distance & bearing) 99
Fig. 5.11. LTE-M & AIS ship domain applied... 100
Fig. 5.12. LTE-M & AIS ship domain applied result (Domain 8.0L, 3.2L) 102
Fig. 5.13. LTE-M & AIS ship domain applied result (Domain 16.0L, 6.4L) 102
Fig. 5.14. LTE-M & AIS ship domain applied result (Domain 24.0L, 9.6L) 103
Fig. 5.15. LTE-M & AIS ship domain applied result (Domain 32.0L, 12.8L) 103
Fig. 5.16. e-Navigation monitor(RICM screen) 104