臺灣博碩士論文加值系統

紀錄船舶真實動態的自動識別系統(Automatic Identification System, AIS)數據是海上交通的強大「大數據」，可作為擁擠航道的風險評估和船舶導航之重要參考資料。本研究提出基於AIS數據的兩種方法：資料擷取閘線法(crossing-line method) 和蒙地卡羅方法(Monte Carlo method)評估船舶航行於進港航道上的安全性。資料擷取閘線法使用來自單一閘線或多條閘線上所擷取的船舶軌跡參數之機率密度函數來評估船舶擱淺機率。蒙地卡羅方法考慮船舶軌跡的因果關係，透過迴歸算法，建立船舶軌跡參數與一條或多條交叉線之間的關係，應用蒙地卡羅法評估擱淺機率。透過分析高雄及基隆港的進港船舶行為，定量評估防波堤口和航道的擱淺風險。從兩種方法獲得的擱淺機率的結果彼此一致，本文提出之方法可促進擁擠的水道和港口的航行安全。
本研究採用臺灣近海區域之船舶AIS資料分析，使用近年42個月之AIS資料，每月總計超過一千五百萬筆資料。於高雄港之案例分析，本研究分別計算以船長分類的三種船舶尺寸之貨櫃船及散裝船於堤口(entrance)以及內航道隘口(pass)的風險度。在本文中將船舶潛在擱淺風險度定義為船體超出了航道邊界的機率或可能造成擱淺之機率，而不是實際擱淺機率。使用潛在擱淺場景機率(probability of the potential grounding scenario, PPGS)表示風險度。PPGS多閘線法用於計算堤口區域以及內航道區域的風險值。得到堤口區域的風險值在10-4量級以下。而內航道區域右舷側風險值較高，在 10-2至10-3量級之間。比較單閘線法與多閘線法，結果顯示兩者一致性高。運用蒙地卡羅法的PPGS得到：貨櫃船堤口的風險值在10-3量級以下，但貨櫃船的Panamax和Post-Panamax型船在內航道右舷的風險值偏高，在10-2量級，明顯高於堤口處。蒙地卡羅法相較於多閘線的結果會大一個量級，但整體的趨勢仍是類似的。
本研究中探討基隆港進港貨櫃船之行為，由於港灣環境不同，基隆港根據靠泊碼頭，船舶尺度和季節條件等因素分類AIS數據，再建立考慮進港船舶橫向位置，速度和艏向的船舶行為模型。依據可能發生危險的區域(東延伸堤頭、堤口和內防波堤)分別計算潛在碰撞機率(PPCS)。分析結果顯示，進港船舶與東防波堤的潛在碰撞風險PPCS相當低(均在 10-4量級以下)，而堤口西側及內防波堤東側的風險會稍高於其他區域，但也不高於 10-4量級，說明這些區域對進港船舶相對安全。
本研究的主要貢獻如下：
1. 提出基於AIS數據以評估船舶進港航行安全性的兩種方法，根據分析實際紀錄真實船舶動態可以獲得船舶操縱行為趨勢，統計分佈結果可以很好地符合閘線上所擷取之統計數據，這些結果構成了統計分析的基礎，適用於評估進港船舶操縱安全性。
2. 本文提出之方法應用於高雄港和基隆港進港船舶風險度評估，碰撞風險的結果與政府官方紀錄一致。從AIS獲得的船舶操縱訊息不僅有助於早期識別船舶導航異常以及海上監視，更有助於開發人工智慧導航的關鍵知識庫。
3. 本方法適用於規畫安全航路，並可提供有關於航路設計以及交通管理的寶貴訊息。本研究的結果，基於AIS數據的交通模式和船舶行為模型，能為港口交通管理單位提供了充分的鑑別知識，替港口當局VTS（船舶交通服務）站的船舶交通監控提供了有價值的資訊，為監管和運輸安全改進提供有用的參考。

The Automatic Identification System (AIS) data, which records the real ship behaviors, is a powerful “Big Data” provides an important reference for risk assessment and ship navigation for maritime traffic. This study proposes two methods based on AIS, namely, the crossing-line and the Monte Carlo methods, are used for navigation safety assessment in approaching channel. The crossing-line method evaluated the ship grounding probability using the probability density functions of the ship track parameters from single-crossing-line or multi-crossing-lines. The Monte Carlo method considers the causation of the track parameters. Such causation was realized using a regression algorithm by which the relationships among the ship track parameters can be established through a crossing-line with the factors corresponding to the previous crossing-lines. With the relationships, Monte Carlo simulation of ship trajectory evaluation was then applied to assess the grounding frequency. The inbound ship behaviors of Kaohsiung and Keelung port are analyzed to quantitatively assess the grounding risk. The results of the grounding probability acquired from both methods are consistent with each other. The method proposed in this study can be further applied to risk assessment and will promote the safety of navigation in congested waterways and ports.
The AIS data used in the present study spans over a period of 42 months with more than fifteen million datasets per month. In Kaohsiung port case, two types of vessel, container ships and bulk carriers with three ship sizes classified by ship length, respectively are considered to evaluate risk value at entrance and pass. In this study, the potential grounding risk is defined as the probability of the ship's position exceeding the channel borderline that has caused or could have caused grounding accidents, rather than that of the actual grounding events. Using the terminology of "probability of the potential grounding scenario(PPGS)" to denote the risk value. The PPGS multi-crossing-line method of the inbound ships at the entrance are around 10-4 order of magnitude or less. The grounding probabilities at the starboard boundary of the inner traffic lane are 10-2 to 10-3 order of magnitude and are significantly greater than those of the port boundary. From the results of the PPGS at the pass (by single-crossing-line method) and the inner traffic lane (by multi-crossing-line method), both the starboard boundary is more dangerous than the port boundary. Using the Monte Carlo method of PPGS, the following results are obtained: the PPGS of the container ships at the entrance are 10-3 order of magnitude or less. However, the grounding probabilities of Panamax and Post-Panamax ships at the starboard boundary of the inner traffic lane are 10-2 order of magnitude, which is significantly higher than the entrance. The comparison of the results between the Monte Carlo method and multi-crossing-line method shows that the outcome of the Monte Carlo method one order of magnitude is larger than the multi-crossing-line method. However, a similar tendency and acceptable agreement of grounding risk are reached.
Due to the different environment, Keelung Port classifies AIS data according to factors such as berthing dock, ship size and seasonal conditions, and then establishes a ship behavior model that considers the lateral position, speed and heading of incoming ships. PPCS are calculated separately based on the areas where the hazard may occur(head of the eastern breakwater, entrance, and inner breakwater)It shows that most of the PPCS of the inbound ships are less than or equal to 10-4 order of magnitude. It indicates the colliding risk is not significant in the area of Keelung Port. It was a higher PPCS value on the west breakwater at the entrance, but not higher than 10-4 order of magnitude, which suggests that these regions are relatively safe for inbound ships.
The main contributions of this study are follows:
1. Two methods based on AIS data to assess the safety of the inbound navigation of the ship are proposed. According to the analysis of the actual recorded real ship dynamics, the trend of the ship's maneuvering behavior can be obtained. The statistical distribution functions can be fitted well with the data collected on the crossing-lines. These results build up the basis of statistical analysis, which can be applied to evaluate the navigation safety.
2. The method proposed in this study is applied to the risk assessment of ships entering the ports of Kaohsiung and Keelung. The results of the colliding risk are also consistent with the records of the government departments. The navigation information obtained from the AIS data is not only useful for early identification of ship navigation anomalies and maritime surveillance but also for developing a critical knowledge base for intelligent navigation.
3. This risk assessment methodology forms a useful tool in finding safety navigating routes and provides valuable information on waterway designs and traffic management. The results of the present study, the traffic pattern and ship behaviour model based on the AIS data, provide sufficient discriminating knowledge for the Vessel Traffic Service (VTS) station of Keelung Port and can be the useful references for regulatory and transportation safety improvement.

摘要 I
Abstract III
Contents V
Chapter 1 Introduction 1
1.1 Motivation 1
1.2 Goal and Objectives 3
1.3 Expected Contributions 3
Chapter 2 Literature Review 5
Chapter 3 AIS Data Collection and Processing 10
3.1 AIS Data Collection and Processing 10
3.2 Statistical Analysis of Route Data on Crossing-lines 11
3.2.1 Spatial distribution 11
3.2.2 Speed distribution 12
3.2.3 Heading angle distribution 12
3.3 Risk Assessment 12
3.3.1 Probability of the potential grounding/contact scenario (PPGS/PPCS) 13
3.3.2 Probability of grounding/contact (PG/PC) 19
3.3.3 The risk frequency ranking index 20
Chapter 4 Kaohsiung Port Statistical Analysis and Risk Assessment 22
4.1 Analyzed Area - Kaohsiung port 22
4.2 Statistical Analysis of Route Data on Crossing-lines 25
4.2.1 Spatial distribution 25
4.2.2 Speed distribution 27
4.2.3 Heading angle distribution 29
4.3 Statistical Analysis Along Crossing-line 30
4.3.1 Average lateral position 30
4.3.2 Average ship speed 32
4.3.3 Average heading angle 33
4.4 Risk Assessment 35
4.4.1 Probability of the potential grounding scenario (PPGS) 35
4.4.2 Probability of grounding (PG) 43
4.4.3 The AIS sensitivity test 46
4.4.4 The Reliability test 48
Chapter 5 Keelung Port Statistical Analysis and Risk Assessment 53
5.1 Analyzed Area - Keelung port 53
5.2 AIS Data 55
5.3 Statistical Analysis Along the Inbound Fairway 57
5.4 Statistical Analysis of Route Data on Crossing-lines 66
5.5 Crossing-line Method for the Contact Risk 75
5.5.1 Single-variate analysis for the PPCS 75
5.5.2 Two-variate analysis for the PC 78
5.6 The Risk Chart 82
Chapter 6 Conclusion 84
References 87

[1] Goerlandt F, Montewka J. Maritime transportation risk analysis: Review and analysis in light of some foundational issues. Reliability Engineering & System Safety. 2015;138:115–134. doi:10.1016/j.ress.2015.01.025
[2] Quy N, Vrijling J, Gelder PHAJM, Groenveld R. On the Assessment of Ship Grounding Risk in Restricted Channels. Varna, Bulgaria; 2006. p. 25–27.
[3] Martins MR, Maturana MC. Human error contribution in collision and grounding of oil tankers. Risk Analysis: An Official Publication of the Society for Risk Analysis. 2010;30(4):674–698. doi:10.1111/j.1539-6924.2010.01392.x
[4] Akhtar MJ, Utne IB. Human fatigue’s effect on the risk of maritime groundings – A Bayesian Network modeling approach. Safety Science. 2014;62:427–440. doi:10.1016/j.ssci.2013.10.002
[5] Mazaheri A, Montewka J, Kujala P. Modeling the risk of ship grounding—a literature review from a risk management perspective. WMU Journal of Maritime Affairs. 2014;13(2):269–297. doi:10.1007/s13437-013-0056-3
[6] Pietrzykowski Z, Gucma L. Theoretical basis of the probalilistic-fuzzy method for assessment of dangerous situation of a ship manoeuvring in a restricted area. Annual of Navigation. 2001;(3):111–125.
[7] Gucma L. The risk assessment of ships manoeuvring on the waterways based on generalised simulation data. In: WIT Transactions on The Built Environment. Vol. 94. 2007. p. 411–418. doi:10.2495/SAFE070411
[8] Quy N-M, Vrijling J-K, vanGelder P-H-A-J-M, Groenveld R. Methods to assess safety criteria of approach channels with respect to the acceptability of ship grounding risks. In: The 6th International Symposium on Navigation, Gydnia, Poland. 2005. p. 69–76.
[9] Lan J, van Doorn JTM, ten Hove D. Probabilistic design of channel widths. PIANC MMX Congress, Liverpool, UK. 2010.
[10] Gucma L, Montewka J. Landborne laser rangefinder measurements for navigation safety assessment. European Journal of Navigation. 2005;3:1–6.
[11] Zalewski P, Montewka J. Navigation safety assessment in an entrance channel, based on real experiments. In: International Congress of the International Maritime Association of the Mediterranean (IMAM). Varna, Bulgaria; 2008. p. 1113–1117. https://research.aalto.fi/en/publications/navigation-safety-assessment-in-an-entrance-channel-based-on-real
[12] Mascaro S, Nicholso AE, Korb KB. Anomaly detection in vessel tracks using Bayesian networks. International Journal of Approximate Reasoning. 2014;55(1, Part 1):84–98. (Applications of Bayesian Networks). doi:10.1016/j.ijar.2013.03.012
[13] Montewka J, Hinz T, Kujala P, Matusiak J. Probability modelling of vessel collisions. Reliability Engineering & System Safety. 2010;95(5):573–589. doi:10.1016/j.ress.2010.01.009
[14] Goerlandt F, Kujala P. Traffic simulation based ship collision probability modeling. Reliability Engineering & System Safety. 2011;96(1):91–107. (Special Issue on Safecomp 2008). doi:10.1016/j.ress.2010.09.003
[15] Pratiwi E, Artana KB, Dinariyana A a. B. Fuzzy Inference System for Determining Collision Risk of Ship in Madura Strait Using Automatic Identification System. International Journal of Marine and Environmental Sciences. 2017;11(2):401–405.
[16] Rong H, Teixeira A, Soares CG. Evaluation of near-collisions in the Tagus River Estuary using a marine traffic simulation model. Zeszyty Naukowe Akademii Morskiej w Szczecinie. 2015;(nr 43 (115)):68–78.
[17] Eide MS, Endresen Ø, Brett PO, Ervik JL, Røang K. Intelligent ship traffic monitoring for oil spill prevention: Risk based decision support building on AIS. Marine Pollution Bulletin. 2007;54(2):145–148. doi:10.1016/j.marpolbul.2006.11.004
[18] Sawano N, Hamada S, Arola T. Analysis of vessel traffic and safety assessment of the Soya Strait. 2011. p. 361–373. doi:10.2495/SAFE110321
[19] Wang Y, Zhang J, Chen X, Chu X, Yan X. A spatial–temporal forensic analysis for inland–water ship collisions using AIS data. Safety Science. 2013;57:187–202. doi:10.1016/j.ssci.2013.02.006
[20] Talavera A, Aguasca R, Galván B, Cacereño A. Application of Dempster–Shafer theory for the quantification and propagation of the uncertainty caused by the use of AIS data. Reliability Engineering & System Safety. 2013;111:95–105. doi:10.1016/j.ress.2012.10.007
[21] Zhang W, Goerlandt F, Montewka J, Kujala P. A method for detecting possible near miss ship collisions from AIS data. Ocean Engineering. 2015;107:60–69. doi:10.1016/j.oceaneng.2015.07.046
[22] Zaman MB, Kobayashi E, Wakabayashi N, Maimun A. Risk of Navigation for Marine Traffic in the Malacca Strait Using AIS. Procedia Earth and Planetary Science. 2015;14:33–40. (The 2nd International Seminar on Ocean and Coastal Engineering, Environment and Natural Disaster Management, 2014). doi:10.1016/j.proeps.2015.07.082
[23] Zaman MB, Kobayashi E, Wakabayashi N, Maimun A. Development of Risk Based Collision (RBC) Model for Tanker Ship Using AIS Data in the Malacca Straits. Procedia Earth and Planetary Science. 2015;14:128–135. (The 2nd International Seminar on Ocean and Coastal Engineering, Environment and Natural Disaster Management, 2014). doi:10.1016/j.proeps.2015.07.093
[24] Xiao F, Ligteringen H, van Gulijk C, Ale B. Artificial Force Fields for Multi-agent Simulations of Maritime Traffic: A Case Study of Chinese Waterway. Procedia Engineering. 2012;45:807–814. (2012 International Symposium on Safety Science and Technology). doi:10.1016/j.proeng.2012.08.243
[25] Xiao F, Ligteringen H, van Gulijk C, Ale B. Comparison study on AIS data of ship traffic behavior. Ocean Engineering. 2015;95:84–93. doi:10.1016/j.oceaneng.2014.11.020
[26] Mazaheri A, Montewka J, Kotilainen P, Sormunen O-VE, Kujala P. Assessing Grounding Frequency using Ship Traffic and Waterway Complexity. The Journal of Navigation. 2015;68(1):89–106. doi:10.1017/S0373463314000502
[27] Sang L, Wall A, Mao Z, Yan X, Wang J. A novel method for restoring the trajectory of the inland waterway ship by using AIS data. Ocean Engineering. 2015;110:183–194. doi:10.1016/j.oceaneng.2015.10.021
[28] Wu X, Mehta AL, Zaloom VA, Craig BN. Analysis of waterway transportation in Southeast Texas waterway based on AIS data. Ocean Engineering. 2016;121:196–209. doi:10.1016/j.oceaneng.2016.05.012
[29] Pallotta G, Vespe M, Bryan K. Vessel Pattern Knowledge Discovery from AIS Data: A Framework for Anomaly Detection and Route Prediction. Entropy. 2013;15(6):2218–2245. doi:10.3390/e15062218
[30] Daranda A. Neural Network Approach to Predict Marine Traffic. Baltic Journal of Modern Computing. 2016;4(3):483–495.
[31] Yan R, Wang S. Study of Data-Driven Methods for Vessel Anomaly Detection Based on AIS Data. 2019. p. 29–37. doi:10.1007/978-981-13-8683-1_4
[32] Xiao Z, Fu X, Zhang L, Goh RSM. Traffic Pattern Mining and Forecasting Technologies in Maritime Traffic Service Networks: A Comprehensive Survey. IEEE Transactions on Intelligent Transportation Systems. 2020;21(5):1796–1825. doi:10.1109/TITS.2019.2908191
[33] Murray B, Perera LP. An AIS-based deep learning framework for regional ship behavior prediction. Reliability Engineering & System Safety. 2021;215:107819. doi:10.1016/j.ress.2021.107819
[34] Montewka J, Krata P, Goerlandt F, Mazaheri A, Kujala P. Marine traffic risk modelling - An innovative approach and a case study. Proceedings of the Institution of Mechanical Engineers, Part O: Journal of Risk and Reliability. 2011;225:307–322. doi:10.1177/1748006X11399988
[35] Burgers A, Kok M. THE STATISTICAL ANALYSIS OF SHIP MANOEUVRING SIMULATOR RESULTS FOR FAIRWAY DESIGN BASED ON THE INTERDEPENDENCY OF FAIRWAY CROSS-SECTION TRANSITS. Vol. 1. 1988. https://trid.trb.org/view/262871
[36] International Maritime Organization. REVISED GUIDELINES FOR FORMAL SAFETY ASSESSMENT (FSA) FOR USE IN THE IMO RULE-MAKING PROCESS. 2018.
[37] Chiu Y-F, Su C-H, Lee, C-Y, Tsai L-H, Liaw C-T, Chiang M-L, Wei C-H, Luo G-S, Fu Y-C, Chen C-Y. Annual Statistic Report of Oceanographical Observation Data in Kao-Hsiung, Offshore Region at 2016. Institute of Transportation, MOTC; 2017.
[38] Goerlandt F, Kujala P. On the reliability and validity of ship–ship collision risk analysis in light of different perspectives on risk. Safety Science. 2014;62:348–365. doi:10.1016/j.ssci.2013.09.010
[39] Chiu Y-F, Su C-H, Lee C-Y, Tsai L-H, Liaw C-T, Chiang M-L, Wei C-H, Luo G-S, Fu Y-C, Chen C-Y. Annual Statistic Report of Oceanographical Observation Data in Keelung Offshore Region in 2016. Institute of Transportation, MOTC; 2017.
[40] Keelung Harbor Bureau. The Keelung Harbor Vessel Traffic Service Manual. English Version 1.0. Keelung, Taiwan; 2011.
[41] Huang J-C, Nieh C-Y, Kuo H-C. Risk assessment of ships maneuvering in an approaching channel based on AIS data. Ocean Engineering. 2019;173:399–414. doi:10.1016/j.oceaneng.2018.12.058

[42] Nieh C-Y, Lee M-C, Huang J-C, Kuo H-C. Risk assessment and traffic behaviour evaluation of inbound ships in keelung harbour based on ais data. Journal of Marine Science and Technology. 2019;27(4).
https://jmstt.ntou.edu.tw/journal/vol27/iss4/2. doi:10.6119/JMST.201908_27(4).0002