Return to Session D2 Next Abstract

Session D2: Marine Vehicle Navigation

A Route Plan Technique with Risk Contour for Autonomous Navigation of Ships Carrying HNS
Mingi Jeong, Korea Maritime and Ocean University, South Korea; Moonjin Lee, Korea Research Institute of Ships & Ocean Engineering(KIOST)/Korean Society of Maritime Environment & Safety, South Korea; Eun-Bang Lee, Korea Maritime and Ocean University/Korean Society of Maritime Environment & Safety, South Korea
Location: Spyglass

As many ships carrying hazardous and noxious substances (HNS) have been operated all over the world, and cutting-edge technologies using Internet of Things (IoT), big data, and Artificial Intelligence (AI) are rapidly developing, it is expected that autonomous HNS vessels will be in service in the near future. Thus, it becomes essential to safely operate autonomous HNS vessels by preventing marine accidents arising from risks existing at sea. To ensure the safe navigation of autonomous ships, a totally new type and method of route plan technique is required, totally different from an empirical and qualitative method, which is regarded as traditional. With the traditional method, a navigating officer intuitively and qualitatively decides how the ship should navigate along a route planned on a nautical chart. For example, the officer in charge of route planning just followed senior officer’s order or selected the same route which used to be taken by the former officers without any further detailed verification. In addition, it appears that the officer does not have a specific standard to determine at what distance the ship is supposed to safely pass by obstructions, shorelines, and other hazards in the corresponding circumstance. Many marine accidents resulting from human errors still occur because the safety standard for route planning differs among individuals.
Therefore, in order to resolve this kind of problem, a route plan technique using risk contour derived from quantitative computation, which makes it possible to utilize real-time data at sea, is proposed and developed in this paper. Moreover, the route plan technique based on risk contour concerns not only the safety of the navigational route, but also efficiency in comparison with the distance and the number of waypoints.

With regard to the methodology, as most topographical maps or nautical charts are represented in contour in order to describe and visualize the altitude of lands or the depth of waters, the risk contour is suggested so that the navigating officer can easily understand the degree of risks. The contour is visualized by connecting the areas that have the same risk index. Thus, the contour can help the officer to plan and determine where to navigate and how to avoid dangers, based on the quantitative approach.
Although there are many risks associated with HNS marine accidents, the scope of this study was limited to navigational traffic risks concerned with three types of accidents: collision, grounding, and touching. The hazards existing along the ship’s route and affecting the risks are defined as static hazards, such as wrecks, rocks, obstructions, islets, and so on. Accordingly, dynamic mobile objects, including other ships, were not considered.
To evaluate navigational traffic risks quantitatively, the risks are calculated in combination with hazards, corresponding vulnerability, and arising impacts. Next, each factor belonging to the hazards, vulnerability, and impacts was identified based on a review of previous studies and other technical publications. Those factors were again conceptualized and structured to derive risk indexes based on the evaluation matrix. The matrix was adopted with reference to external sources, and technical discussions with experts in the field. Finally, in order to gain the total sum of each risk calculated from each hazard, we determined weight coefficients by reflecting the statistical data of maritime accidents, results from technical discussions with experts, and responses to analytic hierarchy process questionnaires.
The unit area where the risk is evaluated is circular. The concept of a unit circular area was introduced by the ship’s position fixing interval. In other words, as the position fixing interval ensures that the ship will not encounter any hazards during the specific period of time, we can assess the risks based on the circular areas made by the interval, as per the ship’s navigational safety. Additionally, the radius of a unit circle is the distance equal to the multiplication of the interval and the ship’s transit speed.
The procedure for making the route plan in accordance with risk contour is algorithmic. First, the departure and arrival point of the ship is designated. Next, risk evaluation is conducted within each area with hazards. Afterwards, No Go Areas, places like shores where the ship is never to approach, are identified. The No Go Areas can vary in accordance with the target ship’s speed and water depth in consideration of the ship’s draft. Next, the areas with the same risk values are connected and visualized by contour line, and we can derive a navigable width of the ship within the region based on the user setting of the allowable risk value; In other words, the region containing the No Go Areas as well as the region where the risks are above the allowable value. Within the navigable area, diverse course options could be given in consideration of the safety related to the risks, and the efficiency based on distance of course and the number of waypoints. Finally, the route is optimized as an accumulation of risk-based course legs. If any data affecting risk factors changes, the evaluation will go to feedbacks in order to reflect the change.

As for the result of this study, prior to applying the method to an actual route plan, we conducted simulations of a model liquefied natural gas (LNG) ship on a model chart where hazards were randomly distributed, and the direction which the ship was heading to was also given. The risk indexes were computed in accordance with the above methodology, and within the navigable area derived by risk contour lines, it was possible to choose diverse course options based on the safety. In addition, we compared the distance of each course, the number of waypoints and the average risk along the course in order to optimize the course in the region. It was meaningful that probable navigable width was suggested with the use of the risk contour, and that the optimized route could be decided by the user’s setting of the allowable risk, and efficiency concerning the distance together with the number of waypoints.
After the simulation on the model chart, verification was carried out to compare the route plan technique with risk contour and the manual route plan made by the officers. We conducted surveys which were intended for the officers, including captains, who have or had experience in planning a route. This verification was to have them plan a route based on safety and efficiency in the designated area of the given chart, and compare the results. Another verification was made to actually monitor and track several ships transiting the same designated area. To analyze the ships’ courses, we used real-time information of Automatic Identification System (AIS). The two practical methods for verification are important because of the application of the risk-based route plan made by the quantitative risk contour. It was found out that the optimized route plans by the technique with risk contour showed the results similar to the surveys of experienced officers and the actual movements of ships in the designated are.

In conclusion, as autonomous ships are expected to emerge in the real-world shipping industry in the near future, and the quantity of HNS cargoes transported by the ships is expected to increase, it is inevitable that the next step is to consider an innovative navigating method. Therefore, the route decision technique using risk contour was applied in order to operate the ships using a quantitative and objective approach, a technique different from the empirically-based route plan which inherently has human errors. First, the concept of the risk contour was introduced to develop the route plan technique. Second, the area where the risk is determined was defined. Next, the factors related to the navigational traffic risk of HNS ships were identified and structured for risk assessment. Moreover, the risks which were calculated in the unit circular areas were visualized as a contour map suitable for determining the navigable area in the region. Finally, the route was optimized among several options within the navigable area in consideration of both safety and efficiency.
This study is very meaningful because the proposed method makes it possible to assess the navigational traffic risk and plan the route based on quantitative analysis, which counteracts the deficiency of the qualitative and empirical method, and prevents marine accidents resulting from human errors. Thus, it has a role in deciding the risk-based route by increasing the understanding of the risk through the visualized contour map. In addition, this technique will have more benefits, as the new technology such as Electronic Chart Display and Information System (ECDIS), which does not require much human effort. Furthermore, the risk-based route decision will be useful and advantageous not only for autonomous ships once the algorithm realizes the fully automatic system, but also for current vessels already operating at sea, because it can be applied to the route decision considering the navigational risks.
One of the elements to be studied in the future will be the enhancement and improvement of the technique reflecting real-time data including dynamic factors like other ships through the connection with navigational equipment in order to realize autonomous navigation containing automatic maneuvering.

Return to Session D2 Next Abstract