WO2011027037A1

WO2011027037A1 - Intelligent waterway risk indication system and a related method

Info

Publication number: WO2011027037A1
Application number: PCT/FI2010/050688
Authority: WO
Inventors: Tony Rosqvist
Original assignee: Valtion Teknillinen Tutkimuskeskus
Priority date: 2009-09-04
Filing date: 2010-09-06
Publication date: 2011-03-10
Also published as: FI20095914A0

Abstract

A risk indication system (400) for providing an incident probability indication for a collision or grounding in maritime traffic is provided. The risk indication system (400) is configured to obtain data related to maritime traffic. The data may comprise identification information, position, course, and speed information of one or more vessels. The system is configured to perform an incident probability calculation by utilizing the data related to maritime traffic, and to generate an indication of the incident probability.

Description

INTELLIGENT WATERWAY RISK INDICATION SYSTEM AND A RELATED METHOD

FIELD OF THE INVENTION Generally the invention relates to maritime safety. Particularly, however not exclusively, the invention relates to a statistical prediction on a grid.

BACKGROUND

Several types of vessels for different purposes are sailing at sea. Passenger ships operate between harbors carrying thousands of passengers in every trip, cruise ships cruise tourists just for pleasure voyages and yachting is common leisure activity. However, marine cargos shipped at sea are a different story; cargo vessels and tankers transport millions of tons of cargos and oil at sea all over the world every year. So, safety and efficiency are main issues in maritime traffic since the amount of the traffic is enormous and, for example, in the Gulf of Finland, e.g. oil tanker traffic is still increasing. Avoiding maritime accidents is crucial for saving lives, of course, but also for avoiding catastrophic environmental pollution that an oil tanker accident would cause. Direct and indirect costs of that kind of accident would easily climb in millions of Euros. Nowadays, a vessel traffic service (VTS) is widely used for monitoring waterway traffic. VTS is a monitoring system established by maritime safety or port authorities and it is designed to improve safety and efficiency of navigation in a limited geographical area by keeping track on vessel movements. The VTS traffic image is compiled and collected by means of advanced sensors such as radar, AIS (Automat- ic Identification System), direction finding, CCTV (Closed-Circuit Television) and VHF (Very High Frequency) radiotelephony or other co-operative systems and services. An example of a VTS traffic image is presented in Fig. 1. Vessels equipped with AIS between Helsinki and Tallinn are presented in real time in the figure with different colors' representing different vessel types. Fig. 2 illustrates AIS system and data send by it. The AIS is a short range coastal tracking system used on ships and by VTS for identifying and locating vessels by electronically exchanging data with other nearby ships and VTS stations. Informa- tion such as unique identification information, position, course, and speed can be displayed on a screen or an Electronic Chart Display and Information System (ECDIS). In addition to the above mentioned essential information, an AIS message can contain other information, such as length, breadth and type of vessel, time stamp, draft and cargo information, navigational status, e.g. anchored, not under command, fishing, destination and time of arrival, for example. An example of an AIS track graph is illustrated in Fig. 3. AIS position information transmitted by vessels is compiled during four days and the result illustrated as a graph. The route of vessels during the time period forms a line in the graph. Busy fairways and harbors can be clearly seen in the graph, since there are thick lines to indicate those. In the graph it is clearly seen that vessel are tending to particular routes so that one route is heading to the East and one to the West.

AIS is intended to assist the vessel's officers of watch and allow maritime authorities to track and monitor vessel movements, and integrates a standardized VHF transceiver system such as a LO AN-C or Global Positioning System receiver, with other electronic navigation sensors, such as a gyrocompass or rate of turn indicator. In practice, AIS is required to be fitted aboard international voyaging ships with gross tonnage (GT) of 300 or more tons, and all passenger ships regardless of size. It is estimated that more than 40,000 ships currently carry AIS class A equipment. A comprehensive traffic image is formed of all factors influencing the traffic as well as information about all participating vessels and their intentions. By means of the traffic image, situations that are developing are evaluated by VTS personnel. The data evaluation depends to a great extent on the quality of the data that is collected and the ability of the operator to combine this with an actual or developing situa- tion. Because the monitoring and the evaluation of the maritime traffic are performed by the VTS personnel, there is always an influence of a human factor and the possibility of a human error.

Prior art systems know several early warning system for maritime traffic, but these systems typically use straight forward vector calculus and piecewise linear extrapo- lation. Unfortunately, these systems take into account neither natural unsteadiness nor trend changes in trajectories, which basically make them suitable for monitoring solely fairway traffic. SUMMARY OF THE INVENTION

The objective of the embodiments of the present invention is to at least alleviate one or more of the aforesaid drawbacks evident in the prior art systems in the context of maritime safety. The objective is generally met by a risk indication system and a related method in accordance of the present invention. The system and the related method are used to learn and predict the movement of vessels in order to prevent collision or grounding in maritime traffic and/or observing the deviation from a route. The method of continuous learning utilizes a dynamic computation algorithm for calculating probabili- ties about future vessel positions in order to recognizing possible incidence zones of vessels moving on a two-dimensional surface, in where position data projected in time, is used in current time, to update incident zones and incident probabilities.

A risk indication system for providing an incident probability indication of a collision or grounding in maritime traffic is characterized by the features disclosed in a corresponding system claim 1.

A method for providing an incident probability indication of a collision or grounding in maritime traffic is characterized by the features disclosed in a corresponding method claim 9.

The utility of the present invention arises from a plurality of issues. The risk indica- tion system of the present invention may improve the efficiency and reliability of, for example, VTS accident prevention by means of automatic monitoring, identification, classification and alarming of the risk of collision or grounding on a large sea area. In other words, the risk indication system may assist e.g. VTS personnel to evaluate maritime traffic situations and recognize developing risks of collision and grounding. The system may also be used as a standalone system, for example, in a vessel. Some embodiment of the system may eliminate collision or grounding of vessel(s) on condition that e.g. VTS personnel and officers of watch in the vessel(s) react to indicate risk development by the arrangement. The risk indication system of the present invention may also be used for observing the deviation of a vessel from the route, which may indicate malfunction of steering systems, e.g. power failure, or in some case even hijacking of the vessel.

Furthermore, the risk indication system could be extendable to support an expert system that recommends the navigator or the VTS operator the safest evasive action based on the dynamics of the situation. Prior art solution using vector calculus and piecewise linear extrapolation are mainly suitable for monitoring fairway traffic. The present invention, however, enables the calculation of prediction data of totally non-linear development and takes into account both natural unsteadiness and trend changes in trajectories, which can be seen in probability distribution of an incident zone and continuous relearning of parameters.

The expression "vessel" may refer herein to passenger vessels, cargo vessels, tankers, high speed crafts, tugs, pilot vessels, yachts and other maritime vessels. The vessels advantageously contain equipment for sending at least identification infor- mation as well as position, speed and course data.

The expression "course data" refers herein the direction information of a vessel.

Next, the expression "maritime traffic" refers herein to waterway traffic of vessels at sea or in some other water area capable for boat traffic.

The expression "incident zone" may refer herein to an area, where a collision or grounding of vessel/vessels may occur or where the distance(s) between two or more vessels is/are too short in accordance of maritime safety regulations.

Further, the expression "data related to vessels in maritime traffic" may refer herein to the information a vessel transmits about its state via one or more known transmission system. Similarly, the expression "data related to maritime traffic" may refer herein the information about vessels in maritime traffic the risk indication system obtains, e.g. from database, or by some other ways.

The expression "transmission system" refers herein to systems or arrangements, such as AIS, in a vessel used to exchange data related to the vessel in maritime traf- fic with other nearby ships and VTS stations.

Still, the expression "monitoring system" refers herein to systems or arrangements, such as VTS station, utilized by harbor or port authorities for monitoring maritime traffic.

In addition, the expression "initialization material" refers herein to data related to maritime traffic, which is collected before introducing the risk indication system with desired parameters, such as vessel type, with or without cargo, cargo type, sea- son, weather condition and time of day, used for initializing the risk indication system.

Further, the "learning material" refers herein to data related to maritime traffic.

Various embodiments of the present invention are disclosed in the attached depen- dent claims. The objects of the invention are achieved with the features of claims 1 and 9. Further advantageous features of the invention are described in dependent claims

BRIEF DESCRIPTION OF THE RELATED DRAWINGS

Next, the invention is described in more detail with reference to the appended drawings in which

Fig. 1 illustrates an example of a VTS traffic image.

Fig. 2 illustrates a concept of Automatic Identification System (AIS) and data sent by it. Fig. 3 illustrates an example of an AIS track graph.

Fig. 4 depicts a concept of the present invention in general.

Fig. 5 illustrates the gathering of the position data of particular vessels in particular route.

Fig. 6 illustrates an example of expected routes of vessels and simulated routes produced with Markov model.

Fig. 7 illustrates a flow diagram of an embodiment of a method according to the present invention.

DETAILED DESCRIPTION OF THE EMBODIMENTS Figures 1-3 were already contemplated hereinbefore in connection with the review of the background of the invention. Figure 4 discloses, by way of example only, a sketch of the present invention, wherein an internal of a risk indication system 400 in accordance with the present invention comprises a processor 402, a memory 404, one or more logic blocks 406, a user interface (UI) 408 and, optionally, one or more additional transceivers 410. The logic blocks can be additional, too.

The basis of the present invention is the data related to vessels in maritime traffic. The vessels with transmission system send information about their position and status. The information the vessel may send is listed in the table in Fig. 3. According to an embodiment, identification information, position, course, and speed information of a vessel are used in the present invention. The vessel may also send another information, such as vessel type, cargo type, timestamp and destination information. The data transmitted by vessels may be compiled to a database, which can be part of or, at least, functionally connected to the risk indication system. The data related to maritime traffic is used for calculating risk probabilities for vessels about the real time position, speed and course data, but it is also used as a learning material for the risk indication system.

Generally, the risk indication system 400 can be part of or, at least, functionally connected to 418 a monitoring system 414, such as VTS. Typically, the monitoring system 414 comprises a transceiver or transceivers, which obtain(s) data related to vessels in maritime traffic for example from vessels or base stations 416. On the other hand, the risk indication system can be a standalone system having its own transceiver(s) and possible database for data related to maritime traffic. This kind of standalone system may be located, for example, in a vessel. In the case of standalone system, the display of UI 408 may be used for displaying data related to ves- sels in maritime traffic and e.g. alarm indications. In addition, the standalone system 400 may be implemented without logic blocks 406 and the probability calculation may be performed somewhere else. The calculation services may be provided by e.g. a server of a monitoring system, for example.

While being part of or functionally connected to the monitoring system, the data transmission between base stations or vessels 416 and the monitoring system 414 may also be arranged via the risk indication system 400 with its optional transceivers) 410. In addition, the risk indication system 400 may use a database for data related to maritime traffic of the monitoring system 414 it is connected to or it may have the database of its own. Furthermore, the processor 402 of the risk indication system comprises at least one processor, such as one or more microprocessors, micro-controllers, DSP's (digital signal processor), programmable logic chips, etc., or any desired combination thereof. In addition, the processor 402 may comprise a plurality of cooperating proces- sors or sub-processors. The processor 402 is configured to execute the code stored in the memory 404, which may imply processing instructions and data related to the risk indication system functionalities of the present invention and optionally other functionalities, such as OS related functionalities, I/O-related functionalities, and other applications. Likewise, the processor 402 controls the logic blocks 406. In addition, the memory 404 may be divided between one or more physical memory chips or other memory elements, and it may comprise code, e.g. in a form of a computer program/application for the risk indication system, and other data. Moreover, the memory 404 may further include other storage media, such as a preferably detachable memory card, a floppy disc, a CD-ROM, fixed storage medium such as hard drive. The memory 404 may be non- volatile, e.g. ROM, PROM, EEPROM or flash memory, and/or volatile, e.g. RAM, by nature.

Still, the risk indication system may comprise optional logic blocks 406, such as a math block, for processing data related to a maritime traffic. The logic blocks 406 are typically stored in the memory 404 of the risk indication system and they can be executed by the processor 402. The logic blocks 406, especially math blocks, can be used to, for example, Bayesian statistical estimation utilized in this invention.

The UI (user interface) 408 may comprise a display, or a connector to an external display or data projector, and keyboard/keypad or other applicable control input means (e.g. touch screen or voice control input, or separate keys/buttons/knobs) configured so as to provide the user of the system 400 with practicable data visualization and device control means.

Optionally, the risk indication system 400 may contain one or more the transceivers 410, complying with predetermined wired or wireless technology, for receiving data related to vessels maritime traffic sent by, for example, vessels equipped with AIS devices. However, the data related to maritime traffic may be obtained from some other way; for instance, it may be provided by satellite(s) or radar(s). The optional transceiver 410 can be a device that has both transmitting and receiving capabilities or the transceiver may consist of a separate transmitter and receiver, or the risk indication system 400 may contain only receiver(s). Optional transceiver(s) 410 in the risk indication system 400 may be implemented for wireless network technology typically used in maritime data communications.

It is clear to a skilled person that the risk indication system may, in practice, comprise numerous further functional and/or structural elements for providing various beneficial communication, processing, or other features, whereupon this disclosure is not to be constructed as limiting the presence of potential additional elements in any manner.

At first, the risk indication system has to go through an initialization phase. During the initialization phase data related to maritime traffic is collected from the area the risk indication system is supposed to use as an initialization material for the risk indication system. The initialization material is divided into several probability matrix categories according to various parameters, which depend on embodiments.

Figure 5 illustrates the gathering of the position data of particular vessels in particular routes. The initialization material comprises position, course and speed data of different vessel types in maritime traffic with and without cargo and/or wherein the initialization material further comprises vessel type, circumstance, weather condition and light information. A sufficiently large initialization material is needed for reliable developing and verifying the calculation routine of the risk indication system of the present invention. The data is collected in normal situations, i.e. data of abnormal occasions is removed from the initialization material. For example, if the observed vessel is waiting for some other vessel to pass, there will be several data points in that location and these "noisy" data points are removed from the initialization material. For example, about ten trips via some particular route for each vessel types are needed for the initialization material. However, the initialization material will be bigger if different circumstance, light and weather conditions are included in the material.

The number of vessel types in the initialization material may vary depending on embodiments. Typically, the vessel types may be, for example, passenger vessel and cargo vessel. The data of another kind of vessel can also be included. There may be for example 1-6 typical vessel categories. Types related to vessels may also be divided into different weight categories so that vessels are also categorized according to their weight. The weight categories can be divided, for example 1-6 weight categories. It is also possible to define the categories by some other way. The data is collected with and without cargo, for example for vessels such as O O ships, cargo vessels and tankers, because the weight of the cargo affects the movements of a vessel. It is possible to do that also to passenger vessels, but because they usually do not travel without passenger, the benefit of collecting data of passenger vessels without passenger is minor. In some embodiments, the initialization material is collected in different weather conditions and in different time of year for collecting information, for example, on the impact of the wind and ice to the route of vessels. In some other embodiment the data is also collected separately for a days and nights, because light may also have an effect on the route of a vessel. In summary, the initialization material can be divided into categories according to vessel type, weight category, with or without cargo, season of year, weather condition and time of day.

The controlled initialization material is used for developing the calculation routine of the risk indication system and for verifying the same. In action, the initialization material is entered into probability matrices (transition matrices) before utilized in the risk indication system. The probability matrices are formed on the basis of collected position data which are used for the estimation of the transition probabilities between grid elements. In case such position data is missing or unreliable, these estimates are based on expert judgment, for example, by interviewing captain of ves- sels and other navigation personnel. After that, in the initialization phase the initialization material for the various vessel categories is used for developing and verifying the calculation routine. The initialization material adjusts the values in the probability matrix of each category as described in more detail hereinafter.

Next, Fig. 6 illustrates an example of expected routes of vessels and simulated routes produced with a Markov model. First and second orders Markov model are used to replicate the "true" movements of vessels on the grid. The first vessel moves from point A to point B. the expected route is a straight line 602 from point A to point B. however, the Markov model used on the basis of the initialization material simulates routes 604 and 606 for the first vessel as a true route. Similarly, the ex- pected route for the second vessel from point D to point C is a line 608 and the simulated route performed with Markov model is a route 610. It can be seen in Fig. 6 that the expected incident zone is only in the centre of the grid, but true incident zone might be much wider.

Further, statistical prediction and learning of the risk indication system according to the present invention, is now described in more detail. At first, the two dimensional surface is sliced into a grid, i.e. squares, which cover the area to be monitored. The size of the squares in the grid does not have to be equal. For example, the grid can be finer in the area of routes and crossing traffic and coarser elsewhere. The size of squares is a compromise; the finer the grid, the more number-crunching power is needed. However, large square size increases the inaccuracy of the prediction. The appropriate grid size would be about 2-4 times the monitored vessel size, for example.

The squares are indexed { 1 ,..., N} . A square of the grid is interpreted as a possible state 'X' of a vessel. Also, the state sequence X = < xl, ... , xq > of -step is defined for the vessel. Next, the data is assumed to follow Markov chain. Thus, the probability of the sequence is

where π is the probability matrix and the transition probabilities are

In Bayesian analysis, a prior distribution over the transition probabilities follows the Dirichlet distribution

P (Pik, - , PNk) = D ( _lk, ... , _Nk) = D ( ). A posterior distribution after detecting a state sequence length of -step is

V

= Pj_k , v⁷ Lj=i\ⁿjk T ^ajk) where n^ is the number of detected transitions from state j to state k and a_jk are expert judgments on the number of transitions.

The transition probability of /-step from square j to square k for a vessel 'X' is

^■Ps_qk

Local incident probability R_kof two vessels 'X' and Ύ' in square k, when knowing present positions (x₀, yo), will be

where a predictive period is t≥ q + 1.

Global incident probability R, when two vessels are detected in an area, is

t

R « (1— Pr {incident before time i})Pr {incident at time i} i=q+l

Fig. 7 illustrates a flow diagram of an embodiment of a method according to the present invention. In that embodiment the initialization phase has been performed before the risk indication system is delivered, so that step is excluded from the flow diagram and the method starts calculating and indicating an incident probability of a collision or grounding in maritime traffic at step 700. In some other embodiment the user performs the initialization phase before starting to calculate and indicate the incident probability of a collision or grounding. At step 702, the settings of the risk indication system is adjusted with adjusting parameters, such as grid size, vessel types and filter thresholds. The size and shape of squares in the grid may vary so that the grid may be finer e.g. on routes and crossing of routes. The filter threshold defines the value when an incident probability is indicated. The threshold value may vary according to different vessel type combinations so that, for example, vessel with cargo and/or passenger vessel have lower threshold value than e.g. cargo vessel without cargo, or during nights and/or in ice condition the threshold value may be lower than during day in open water. Preferably, the threshold value may be e.g. 2%-25%, more preferably 5%-15%. Some other parameters may also be used. The other adjusting parameters may be, for example, circumstance and weather condition as well as time of day. These parameters have an effect on selecting and updating probability matrices, for example, ice condition during winter time as well as wind speed and direction may be parameters determining the probability matrices. The adjusting and/or the other parameters may be set by a user or they may be in whole or in part default settings or obtained from elsewhere. For example, weather forecasts may, even automatically, be obtained from services distributing forecasts. Some of the parameters, such as grid size and filter thresholds, could be constant or rarely changed parameters, but weather forecasts should be updated frequently, several times, e.g. 3 or 4 times, per day. At step 704, data related to maritime traffic is obtained from a database, where vessels in the maritime traffic provide data. Vessels in maritime traffic transmit data related a vessel, such as AIS data presented in Figure 2 (see table below), frequently, for example every sixth minute, and the data may be compiled to the database. The database may incorporate in or, at least, be functionally connected to the risk indication system. The amount of data provided by vessels in maritime traffic is enormous; since each vessel with transmission system transmits data related to the vessel several, approximately ten or more, times per hour. Depending on a storage capacity of a monitoring system and/or an embodiment of the risk indication system, the data e.g. of one week period is stored in a database and then replaced with a newer data. Typically, the replacement of the data is performed as a continuous process, but in some embodiments the old data may be removed e.g. every week or every month.

At minimum, vessel's identification information, position, course and speed are needed for statistical prediction and learning of the present system. However, vessel type, cargo type, with or without cargo, timestamp and destination information can also be utilized. For instance, the probability matrices may be defined for each vessel type, cargo type and/or with or without cargo. On the other hand, e.g. vessel type, cargo type and/or with or without cargo information may have an effect on e.g. the filtering of the risk. Destination information may be used for e.g. observing the deviation from the route.

The data related to maritime traffic may be obtained from the database e.g. every sixth minute. Further, the data of the tracked vessels is assumed to be synchronous with each other or the optional timestamp information may be used to synchronize obtained data. At step 706, new values are calculated for probability matrixes by using data related to maritime traffic in the database, which is thus used as a learning material. The calculation is performed by using Bayesian estimation described hereinbefore. Generally, only the probability matrixes are updated, which categories the data related to maritime traffic and/or adjusting parameters fall into. The updating of the proba- bility matrixes with new material, which replaces old material, enables continuous learning of the system. In order to avoid errors and diversions, in some embodiment the effect of middle data is weighted more in the learning material than the effect of old and/or new data. If the data of e.g. one week is stored in the database, the portion of the middle data may be for example 50% and, correspondingly, the portion of old and new data 25% each. The portion of the middle data may also be bigger or smaller. The utility of weighing the middle data more is that the old data may be outdated and new data may contain errors. However, in some other embodiment, new data is weighted more than old data. The portion of old and new data may be e.g. 25% each.

Further, at step 708 a probability matrix for each vessel to be monitored is selected. A probability matrix is chosen for a vessel depending on the grouping of vessels, for example, vessel type, cargo type, with or without cargo, season of year, weather conditions and time of day.

Next, at step 710 a local risk and global risk of a collision or grounding is calculated by using selected probability matrixes and newest position data of each vessel to be monitored. The local risk is a risk of a collision or grounding on a particular square of the grid. Correspondingly, the global risk is the risk for collision or grounding in a sub-area comprised of squares with non-zero local risk. As new position data arrives, these risk measures change, as well as, the related sub-area. The risk calculation, performed by using Bayesian estimation of a Markov chain, has been de- scribed in more detail hereinbefore. Preferably up to e.g. 30 order Markov model can be used, so that a possible risk of collision and/or grounding can be estimate preferably 30 minutes - 2 hours before. However, it is also possible to use some other that 30 order Markov model. For example, in a standalone system, where the system may not have a processor as efficient as processors in control centers, such as VTS, it is possible use e.g. 20 25 order Markov model.

At step 712 it is evaluated if local and/or global risk exceeds threshold value. The threshold value may vary according to the embodiment and/or probability matrix categories. If two probability matrix categories that the local and/or global risk concerns to have different threshold values, the system may be configured to select e.g. the lower value for security's sake. There may be a threshold parameter also for a global risk and the size of the area of the global risk may also be defined by a user. For example, user may define that if e.g. five adjacent squares each have e.g. 35% of incident risk, the global risk exceeds the threshold value.

If local and/or global risk exceeds threshold value, an indication of the increased in- cident probability is generated at step 714. Typically, the risk indication is a warning indication, such as blinking and/or colored square on the map and/or audible signal, delivered to monitoring and controlling personnel and/or personnel in the vessel or vessels the risk is concerned to. In addition, a color chart may be used for indicating the risk level. Generally, in that case the risks under the threshold value are presented with ground color of the map. The colored chart may be divided in e.g. 5-10 risk levels or the risk is presented with sliding color scale from e.g. yellow to red. Alternatively, some other way may also be used to indicate the increased risk. In some embodiment and/or in some probability matrix categories the highest risk level may be e.g. 50% and up. In some embodiments the risk indication system may be configured, in order to avoid collision or grounding of vessel(s), to calculate a steering suggestion, i.e. a suggestion to change the course some particular degrees to some particular direction, for the vessel or the vessels, which the increased risk concern(s) to. In the collision risk situation, the steering suggestion to execute an evasive action is typically calculated for a give-way vessel, i.e. the vessel which is obligated to give way according to maritime regulations, and a stand-on vessel is assumed to keep its course. Preferably, the steering suggestion is calculated so that the evasive vessel does not end up right away in a new risk situation in its new course. If the evasive vessel cannot perform the evasive action without a risk of grounding or a collision with another vessel, the steering suggestion may then be calculated for the stand-on vessel or for both vessels. However, this is an especially exceptional case and the navigators of both vessels, and typically VTS operator as well, are informed of the situation.

The steering suggestion may be assigned for the navigator of the evasive vessel and/or the VTS operator. In addition, the navigator of the stand-on vessel and/or another vessels near may also be informed about the steering suggestion calculated for the give-way vessel. If the steering suggestion is assigned to the VTS operator, he/she can evaluate the suggestion and possibly correct or alter it before giving steering instructions to the vessel(s) in a risk situation. If local and/or global risk does not exceed threshold value, or after step 714, it is detected at step 716 if settings of the risk indication system have changed. If settings have changed, the method forwards to step 702. Else, the method continues at step 704.

In some embodiments the system of the present invention may be configured to in- dicate also deviations of a vessel(s) from the route. The deviation may indicate, for example, malfunction of steering systems, e.g. power failure, or in some case even hijacking of the vessel. In the case of deviation indication, the step 712 is configured to evaluate if the deviation is significant. This evaluation may be based on duration of time the deviation has been last and/or the amount of the deviation. The scope of the invention is determined by the attached claims together with the equivalents thereof. The skilled persons will again appreciate the fact that the explicitly disclosed embodiments were constructed for illustrative purposes only, and the scope will cover further embodiments, embodiment combinations and equivalents that better suit each particular use case of the invention.

Claims

1. A risk indication system (400) for providing an incident probability indication for a collision or a grounding in maritime traffic, wherein a monitored area is sliced into squares in a grid, wherein said risk indication system (400) is configured to obtain data related maritime traffic, wherein the data comprises at least identification information, position, course, and speed information of a vessel, and wherein said system is further configured to per- form incident probability calculation for each of said square by utilizing data related maritime traffic, and wherein said device is configured to generate an indication of the an incident probability that exceed preset threshold value.

2. The risk indication system (400) according to claim 1, wherein the data re- lated to vessels in maritime traffic further contains vessel type, weight category, cargo type, timestamp and destination information, whereupon the system is configured to take into account characters related to these information in performing the incident probability calculation, such as mass inertia.

3. The risk indication system (400) according to any preceding claims, wherein the system is configured to perform an initialization of probability matrixes used for generating the indication of the incident probability by using an initialization material, wherein said initialization material comprises real values and/or expert judgments of position, course, route and/or speed data of different vessel types in maritime traffic with and/or without cargo, and/or wherein the initializing material further comprises weight category, circumstance, weather condition and light information.

4. The risk indication system (400) according to any preceding claims, wherein the system is configured to perform a learning of said risk indication system (400) continuously by updating probability matrixes with learning material of data related to maritime traffic and wherein new data related to maritime traffic replaces old data related to maritime traffic in a database.

5. The risk indication system (400) according to claim 4, wherein the effect of middle data is weighted more in the learning material than the effect of old and/or new data.

6. The risk indication system (400) according to any preceding claims, wherein the incident probability is calculated by using a Bayesian estimation.

7. The risk indication system (400) according to claim 6, wherein said Bayesian estimation follows Markov chain and Dirichlet distribution.

8. The risk indication system (400) according to any preceding claims, wherein said risk indication system is configured to calculate a steering suggestion in order to avoid collision or grounding.

9. A method for providing an incident probability indication for a collision or a grounding in maritime traffic, comprising

-obtaining data related maritime traffic, wherein the data comprises at least identification information, position, course, and speed information of a vessel, -performing an incident probability calculation by utilizing said data related maritime traffic, and

-generating an indication of the incident probability.

10. The method of claim 9, wherein an initialization of the probability matrices used for risk computation and indication is performed by using an initialization ma- terial, wherein said initialization material comprises real values and/or expert judgments of position, course, route and/or speed data of different vessel types in maritime traffic with and/or without cargo, and/or wherein the initializing material further comprises vessel category, circumstance, weather condition and light information.

1 1. The method of any of claims 9-10, wherein the method comprises a step of performing a learning of said risk indication system (400) continuously by updating probability matrixes with learning material of data related to maritime traffic and wherein new data related to maritime traffic replaces old data related to maritime traffic in a database.

12. The method of claim 1 1, wherein the effect of middle data is weighted more in the learning material than the effect of old and/or new data.

13. The method of any of claims 9-12, wherein the incident probability is calculated by using a Bayesian estimation.

14. The method of claim 13, wherein said Bayesian estimation follows Markov chain modeling with uncertain transition probabilities where this uncertainty is modeled by the Dirichlet distribution.

15. A computer program product for providing an incident probability indication for a collision or a grounding in maritime traffic, said computer program product being configured to execute the steps of at least one of the previous method claims 9-14, when said computer program product is run at computer.