Availability and integrity model of Automatic Identification System (AIS) Information

TABLE OF CONTENTS

ABBREVIATIONS AND SYMBOLS

INTRODUCTION

1. DEFINITION OF RESEARCH PROBLEM
1.1. General remarks
1.2. Objectives and thesis dissertation

2. SUMMARY OF THE AIS FUNCTIONING
2.1. Introduction
2.2. Idea of AIS introduction
2.3. Information transmitted by AIS used in the dissertation
2.4. AIS network structure of the Gulf of Gdansk
2.5. AIS malfunctions

3. REVIEW OF AVAILABILITY AND INTEGRITY METHODS AND TOOLS FOR 25 AIS INFORMATION RESEARCH
3.1. Methods and tools for research of AIS information integrity
3.1.1. Application of statistic methods to research AIS service information
3.1.2. Statistical analysis of AIS information in accordance with N.Bailey theory
3.1.3. End-user satisfaction model
3.2. Research methods and tools of AIS information availability
3.2.1. Research method of AIS information availability by A.Hori
3.2.2. Research method (LIC) of AIS availability and coverage area in accordance 30 with Lapinski & Isenor
3.2.3. Research method (HPC) of AIS availability and coverage area in accordance 32 with Hammond & Peters

4. RESEARCH METHODOLOGY FOR ESTABLISHING THE PROBLEM 35 THROUGH AVAILABLE TOOLS AND MODELS
4.1. Preliminary presentation of research models
4.1.1. General remarks
4.1.2. Option 1 - Method for research of AIS information integrity with the use of Fault Tree Analysis (FTA)
4.1.3. Option 2 - Method for research of AIS information availability and integrity 39 with the use of Markov Processes
4.1.3.1. Definitions
4.1.3.2. Stationary distribution
4.2. Remedy research by available models and tools
4.2.1. General remarks
4.2.2. Developing data - "post-processing"
4.2.3. Decoding AIS information
4.2.4. Assumption for availability research of AIS information
4.2.5. Availability structure of AIS information - determination of the object study
4.2.6. Assumption for integrity research of AIS information
4.2.7. Evaluation completeness criteria of AIS information
4.2.8. Completeness structure of AIS information - determination of the object 54 study
4.2.9. Evaluation integrity criteria of AIS information
4.2.9.1. Evaluation integrity criteria of AIS message No
4.2.9.2. Evaluation integrity criteria of AIS message No
4.2.10. Preliminary assumptions of integrity research for AIS information

5. RESEARCH SOLUTION
5.1. Characteristics of input data
5.2. Research outcomes of AIS binary data availability
5.3. Research method of AIS information availability
5.4. Research outcomes of information completeness concerning true heading
5.5. Research method of AIS information completeness concerning true heading
5.6. Research outcomes of information completeness concerning rate of turn
5.7. Research method of AIS information completeness concerning rate of turn

SUMMARY, CONCLUSION, DIRECTION OF FURTHER RESEARCH

BIBLIOGRAPHY

ABBREVIATIONS AND SYMBOLS

illustration not visible in this excerpt

INTRODUCTION

Automatic Identification System (AIS) is a data exchange system which was introduced to improve shipping safety and the possibility of exchanging data, at a country and international level, about ships heading to or from ports, as well as exchanging data relating to passengers and dangerous or environment-polluting cargo carried by ships. The main purpose of introducing AIS was to offer a wider spectrum of available, continuous and reliable navigational data. It became common to use data transmitted through AIS in order to enhance shipping safety. Apart from being useful for traffic control in a marine area, AIS data can be a very important source of information used in collision avoidance process. However, some reservations are voiced with regard to the unconditional reliance on the data transferred through this channel. Thus, it is justifiable to question how and to what extent data coming from AIS is available, complete and reliable. The studies on the technical specification of the system [ITU-R.M.1371, 1998] allow to assume that errors classified as lack of reliability will depend on performance of sensors co-working with AIS and human errors made by operators, and those fitting the devices. The notion of integrity in marine navigation is referred to the radio navigation system. The integrity concept appeared in 2001. According to [FRS, 2001] integrity includes the ability of the system to provide timely warnings to users when the system should not be used for navigation. From 2005 according to [FRP, 2005] integrity is the measure of trust that can be placed in the correctness of the information provided by a navigation system. The referred to document also adds, that integrity assumes the ability of the radio navigation system to provide user with timely warnings when the system should not be used in navigation.

It is questionable to call AIS a ‘navigation system’. Essentially, it is a radio information broadcast channel. For this reason, it does not provide information on data malfunctions. The author considers integrity as the measure of confidence in information received via AIS data. This measure will be expressed by statistical methods.

The information integrity problem had been raised already at the stage of system implementation and some fragmentary outcomes were published in the past. However, it can be assumed that the results from the system introduction period may not represent reality. Thus, it seems reasonable to base the assessment on theoretical analysis, through the development of multi-criteria model describing the AIS information integrity.

AIS information quality analysis published by [Drozd, W., et al., 2006] in A1 maritime area show that up to 30% of information transmitted by ships contain invalid data or is incomplete. Other sources report that only dynamic information coming from devices co-working with AIS can be considered reliable. Voyage related data, which should be input by the navigator is in 80% undefined or defined incorrectly [Bailey N., 2005]. Knowledge of AIS information integrity is fragmentary and it is difficult to recognize it as objective. It should be also taken into account that it comes from the system implementation period. The accelerated implementation of AIS post September 11th 2001 resulted in lack of proper crew training. . This might have been one of the causes of improper system operation in the first years of the twenty-first century.

Information integrity can refer to the information accuracy and could be assessed as position, course or speed error. This error can result from gauge or data transmission quality. The human factor is also important. It affects the occurrence of errors in the transmitted information. AIS information completeness is equally important. AIS information integrity is questioned when the received message contains blank spaces. Thus, the question arises, what tools can provide a full assessment of the information integrity and what is its real integrity.

The measure of system confidence can be information unfitness [Felski A., Jakubowski B., 2004]. AIS information integrity can be described as data sensitive-fault or information unfitness. This is the result of datasets processing [Oleraczuk E., Frydrych A, 1994]. Having established that the matter regards the analysis of the collision avoidance process followed by the watch officer on bridge, valid AIS information components - dynamic, static and voyage related, must be selected. With such approach, analysing all elements of AIS information is not necessary.

Quality of navigational information may be evaluated by the scale of error of processed information expressed in mistake of designate position, course, speed or factors such as human error. The question arises: to what extent information from the AIS is or may be reliable in the future and consequently, what tools can be used to perform such research? Integrity assessment can be based on the assumption that the transmission of navigational information between users can be considered analogously to assess the equipment reliability. If a process of data transfer, resulting in providing a user with complete data at any moment, is a reliable process it can be viewed as a functioning of an information channel and can be investigated with methods included in the theory of reliability as in [Specht C., 2003], [Jakubowski B., 2005], [Felski A., Jakubowski B., 2004], [Specht C., Nowak A., 2005], [Trautenberg H. L., 2005]. Above analysis of the literature allows formulating the following hypothesis.

Hypothesis: There are test methods that can be used to predict the state of the availability and integrity of navigational information provided by AIS.

On the basis of verified hypotheses, determination of data unfitness as a function of time and data processing allows to classify the system to integrity state. Here follows the conclusion that, AIS information is divided into three categories and it is required to investigate them in various methods.

Another problem of system operation is the availability of AIS data transmission channel. Definition of system availability is used for a long time. In 1996, Federal Radio Navigation Plan defines availability as the ability of the system to provide applicable service within the specified coverage area [FRP, 1996]. This definition is repeated in all subsequent editions of this document, most recently in 2008.

AIS service availability is questionable, thus treating AIS as a radio navigation system is debatable. In fact, this is data transmission channel [Felski A., Jaskólski K., 2010]. Data sources come from outside of the system. Potential distortion of information imparted by the system can be considered. However, observations made during the execution of this study allows identification of the availability of AIS with the availability of the data transmission channel.

Then, it is justified to carry out availability research on the basis of received and recorded messages from devices transmitting AIS signals. Given the extraordinarily volatile conditions, even on a limited area (variable number of vessels, their differentiated nature, mutual location, etc.) it is proposed to limit the immutable elements. Number of base stations in the area does not change, their antennas are located at fixed altitudes and are arranged in the same places. Another advantage of these stations in terms of site AIS availability is constant time interval of transmitted messages from these stations.

The investigations related to determining the degree of AIS information integrity should be divided into three main stages.

Stage I:

a. Analysis of the current state of knowledge on the AIS information availability and integrity;
b. Analysis of AIS imperfections indicating availability and integrity limitations of the information system;
c. Selection of components for AIS information integrity research;
d. Development of research methods to define coefficient of information availability and integrity on the basis of the assumptions made.

Stage II:

a. Gathering research material for performing research of system information availability and integrity;
b. Decoding of AIS navigational information based on designed decoding tool;
c. System examination in terms of information unfitness using specialized software (pilot study).

Stage III:

a. Estimating AIS information unfitness on the basis of the criteria adopted for research;
b. Selecting the optimal solution to the problem of the research based on stochastic homogeneous Markov Chains, homogeneous semi-Markov Chains;
c. Determination of quantitative distribution of transitions between different states of the process of AIS information availability and integrity;
d. Prediction of system state of the process of AIS information availability and integrity.

To estimate the availability and integrity of AIS information following test methods were contemplated:

a. Fault Tree Analysis (FTA) - method of analysis designed to determine which type of unfitness, damage of the object, an external event or a combination thereof can generate an object failure. This method is presented in the form of failure tree [Jakubowski B., 2003];
b. Multivariate statistical analysis - method that examines the confidence degree of received
information. Method is based on analysis of recorded data detailing the study of information affecting the safety of navigation. It is possible to determine integrity and availability of binary channel transmission. Multivariate statistical estimation is used in this method;
c. Operating states graph - method of presentation for the reliability structure of the object. This method is used for reliability evaluation. On the basis of stochastic processes, this method is an effective way for reliability estimation of renewable objects [Jakubowski B., 2003];
d. Stochastic methods with the use of Markov Chains - provide a convenient mathematical apparatus enabling the description and investigation of actual random processes. They are an important class of stochastic processes, which allows a mathematical description of the change of random quantities in time.

The advantages and disadvantages of particular solutions are discussed in Chapter 4. In connection with the above a model based on the Theory of the Markov Processes related to Operating Technical Objects was proposed. The characteristics of Markov Chain is that the process state at the moment n + 1 depends exclusively on the moment state n and does not depend on states at previous moments. Furthermore, there is a possibility of short-term prediction of the system after n steps.

AIS, a specific kind of navigation system, which uses radio waves to transmit data with regard to the ship motion parameters, can deliver more accurate information than radar. It can be assumed that it is a data transmission channel from ship sensors. This approach will facilitate research to develop methods for investigating the system information availability and integrity. The navigation characteristics of the system will be determined with use of post-processing method. Not only from the operational point of view, enhanced navigation systems require reliability assessment. Processed navigation information should also be evaluated. A situation may occur, when properly operating system sends out incorrect information that does not fall within the expected limits. Crucial element will be the search for solution in the form of method to examine the AIS information availability and integrity based on the system information and to determine the cause of navigation characteristics limitations.

The AIS availability and integrity assessment is intended to:

a. Develop a method used to investigate the availability of AIS data transmission channel and the associated information integrity;
b. Analyse the information integrity on the basis of proposed method;
c. Process the information derived from devices and to determine the extent to which information is available and reliable;
d. Verify the developed method to estimate availability and integrity coefficients of system.

To solve this task, dissertation is presented in five chapters, as characterized below.

Chapter 1: The research problem was formulated; arguments to prove and dissertation objectives were presented.
Chapter 2: The problem of the functioning of AIS, performance characteristics, principles and purpose of the AIS introduction, the information provided by AIS, the system structure in the Gulf of Gdansk was presented. This chapter ends with the system imperfections presentation.
Chapter 3: Current state of knowledge on research methods for assessing the AIS availability and integrity was presented.
Chapter 4: Description of selected methods of investigating the information availability and integrity has been effected. Choice of methods and tools to solve the problem of research integrity was discussed. The research tool selected by author was characterized.
Chapter 5: Research methodology was presented. Contribution to the development of scientific discipline in terms of system availability and integrity has been presented. Research approach to estimate the characteristics of the navigation system based on selected methodology was presented.

1. DEFINITION OF RESEARCH PROBLEM

1.1. General remarks

According to International Regulations for Preventing Collisions at Sea, Rule V “Look-out” [Rymarz W., 1998] „ Every vessel shall at times maintain a proper look-out by sight and hearing as well as by all available means appropriate in the prevailing circumstances and conditions so as to make a full appraisal of the situation and of the risk of collision”. Effective look-out should ensure early detection of objects. Another task is to assess ship motion parameters and foresee situation development. One of the effective undertakings is radar look-out in any conditions of visibility, especially, when the vessel navigates in area of intensive traffic at night. It is commonly known that Radar and ARPA have some efficacy limitations⁸. Detection of small objects is limited by sea clutters and unfavourable weather conditions (rain, snowstorm) as well as by radio-transmitters operating on similar frequencies. Characteristic phenomenon is the appearance of death-zone (minimum range of detection), radar shadow effect (due to reflections from funnel, mast and other constructions on the vessel) and wave reflection.

As far as ARPA devices are concerned, automatic tracking of objects can be lost during torrential ship manoeuvring or when passing ships within close distance.

In practice, the position information is characterized by a an error not less than about 30 meters, while at higher radar ranges error position may be greater than 100m.

The use of radar information for collision avoidance is effective for constant values of ships movement. Conversely, information delay or information error appears [Śniegocki H., 2002]. The radar information, usually presented in polar coordinates system is subject to some measurement error exceeding approximately 1% of radar range. In practice, the position error is not less than 30 meters, while the position error for navigation radar may be up to 100 m. [IMO Res. MSC.192(79)].

In the paper [T. Stupak, 2012] evaluation of the radar information accuracy derived from radar equipment operating in X- and S-band was carried out. Radar outcomes were compared with AIS information. In 2005-2011 Gdynia Maritime University conducted research in radar laboratory of Navigation Department. The region of the study was the Gulf of Gdansk. Accuracy of radar readings was compared on the basis of anchored objects and aids to navigation. Thus, bearing error 3o and distance error 0,1 nautical mile has been appointed. Bearings and distances to objects have been measured in the time interval of 3 minutes. To everyone's surprise, ARPA dynamic data proved higher accuracy than AIS. Accuracy differences did not exceed the distance of 0.15 nautical miles, while bearing difference was greater than 3o. In the case of dynamic objects, AIS and ARPA indications differed from 0o to 22o. Ships course change of 60o attained indication error up to 38o. The difference in objects speed tracked by AIS and ARPA ranged from 0 to 3.7 knots. In addition, faster information update of course change for AIS targets in relation to the ARPA objects was observed. Therefore it is reasonable to supplement the information derived from radar with additional information delivered automatically and continuously with better accuracy. A system which appears suitable to provide this kind of information is AIS which uses radio waves to transmit data with regard to the ship motion parameters. In addition, AIS provides information on the ship’s name, MMSI and IMO numbers and reduces the number of VHF (Very High Frequency) radiotelephone conversations.

Common use of AIS brings opinions on the imperfections of the system such as lack of transmission or transfer of unreliable information. Studies on the incompleteness and integrity of AIS information published to date are generally linked to the position reports (message no. 1) and static and voyage related data (message no. 5). This problem was analysed, for example by [Harati-Mokhtari A., et al., 2007a], [Bailey N., 2005], [Drozd W., et al., 2007], [Harati-Mokhtari A. et al., 2007b], [Hori A., et al., 2009]. The availability and coverage area problem was considered by [Hori A. et al., 2009], [Lapinski S., Isenor W., 2011], [Hammond T, Peters D., 2012]. The latency of AIS data transmission problem was considered by [Stupak T., 2012]. However, literature studies revealed that there are no publications concerning quality of the dynamic information, particularly the one that is important in the analysis of antiǦcollision manoeuvring. It is very important as the limited information integrity resulting from navigational parameters errors, may be the cause of collision at sea or grounding. It was assumed that the application of stochastic methods allows to determinate the characteristics of the system availability and integrity.

To solve the research problem of AIS information availability and reliability, the system analysis methodology with respect of design constraints and experiment was used. The studies were performed on the basis of recorded messages in txt files. Post-processing method was applied for further work with recorded data. AIS decoder has been developed to perform research. The tool is designed in accordance with [ITU R.M.1371, 1998] and [IEC 61162, 2002]. Furthermore, the author has developed a tool to determine availability and integrity coefficients.

It is crucial to adapt the research tool and conduct reliability analysis in the Gulf of Gdansk area and use methodology that provides reliable research results. It was done through development of research methods for the AIS information availability and integrity based on homogeneous Markov Chains.

Evaluation of AIS information availability and integrity was implemented on the basis of:

a. Research criteria of AIS information reliability and availability on the basis of the system technical specifications;
b. Application of stochastic model based on a homogeneous Markov Chain;
c. Determination of the integrity and availability states distribution of AIS information after n steps;
d. Verification of the proposed research method based on the outcomes.

The development of the method will allow constructing a tool to study the usefulness level of the system information in the aspect of navigation safety and security of shipping. In addition, this method enables a prediction of AIS information integrity and availability.

The proposed research method of AIS information integrity should consider human errors contained in static and voyage related data.

1.2. Objectives and thesis dissertation

The research purpose was:

Development of mathematical model of AIS information integrity and availability based on homogeneous Markov processes.

This problem is the result of lack of forecasting methods of AIS integrity information endearing human and sensors errors.

The realization of the main objective of dissertation is presented in sections:

a. Presentation of the system imperfections and demonstration of existing investigations of AIS information integrity;
b. Development of a decoding methodology for AIS information and a tool for the AIS information unfitness study in accordance with the assumptions adopted for research;
c. Outcomes presentation of AIS unfitness information based on data collected in the Gulf of Gdansk;
d. Application of homogeneous Markov Chains to determine transitions probabilities;
e. The analysis of the system information availability and integrity in the Gulf of Gdansk;
f. Verification of statistical hypotheses to determine the information integrity and availability.

Consistently with the work objective, the theses were formulated as follows: Main thesis:

It is possible to forecast the state of the availability and reliability of the information system on the basis of recorded and processed information extracted from AIS receiver.

Auxiliary theses:

1. There is doubt about AIS information integrity and availability. Method proposed should enable the appointment of system information integrity and availability coefficients.
2. On the basis of information transmitted via AIS base stations it is possible to determine the system state availability.
3. Using Markov processes, it is possible to predict system state availability and integrity after n steps.

2. SUMMARY OF THE AIS FUNCTIONING

2.1. Introduction

AIS is a marine navigation system, commonly used for improving safety at sea. At the time of its introduction the system received negative opinions regarding the correctness of the transmitted information. Reported reservations about the information correctness were related to operator errors, which limited the information integrity.

This chapter presents selected issues of AIS functioning required to explain the research problem. Due to the extensive range of issues, vital information about the system is presented. This information will contribute to development of system messages decoding tool. It is possible to assess the AIS information integrity and availability on the basis of post-processing method.

It is commonly understood that we are able to assess whether data device is reliable, if we can verify the correctness of its operations. However, in case of AIS there is no control over the source of system information (man on-board or unknown device). It should be taken into account that for proper functioning of AIS an efficient communication channel of binary data is required. There is information available about events, when the information obtained through the AIS was incorrect. There is no information about cases when errors are not obvious and probably will be not recognized.

2.2. Idea of AIS introduction

In 1997 a meeting was held at the International Association of Lighthouse Authorities (IALA) in Paris concluding that the preferred option for a new ship-borne identification system should be based on a Self-Organising Time Division Multiple Access (SOTDMA) transmission protocol operating within the VHF marine band. Initially, AIS was envisaged as providing a ship to shore data system that would enhance the operation of Vessel Traffic Services (VTS), providing two-way automatic data communication between ships and port authorities. This had evolved to allow VTS operators to display detailed positional and velocity information from ships at a data rate comparable to radar. The Swedishadvocated SOTDMA system was chosen from a number of other competitive systems, after a number of trials and analysis programmes. Using two 25 kHz VHF marine channels, it offered a reliable system that could operate satisfactorily in the busiest foreseen situation [Norris A., 2007].

IMO members agreed to adopt the SOTDMA system for AIS in preference to a system based on VHF marine-band Digital Selective Calling (DSC). DSC equipment was already mandatory for ships as part of the Global Maritime Distress and Safety System (GMDSS). However, DSC was not capable of transferring dynamic data about ships at rates required for real-time positional and velocity data supplement radar information for collision avoidance purposes. The new system was termed Universal AIS (UAIS) in recognition that the chosen solution also embodied a parallel DSC receiver to allow original mode of use [Norris A., 2007].

AIS quickly became a valuable source of vessel traffic information. Its introduction has reversed the previous problematic situation of scarcity of precise data from ship traffic and instead posed the reverse challenge of coping with an overabundance of data. The number of time-series available for ship traffic and manoeuvring analysis increased from tens, or hundreds, to several thousands. Sifting through this data manually, either to find features of traffic or to provide statistical distributions of decision variables are an extremely time consuming procedures. Applying computer computing techniques together with AIS allowed solving the research problem. The method allowed the study of vessel traffic and simultaneously modelling fairways and traffic separation schemes [Aarsaether K.G, Moan T., 2009]. Transport of hazardous cargo involves a risk of substances entering into the aquatic environment. Therefore, even the smallest accident could cause an ecological disaster. The increase in sea freight number, as well as concerns about pollution of the marine environment pose a significant interest in the optimization of fairways and traffic separation schemes. Vessel traffic analysis is hindered by lack of data and the difficulty of obtaining data from all vessels passing through an area during a time span sufficient to produce statistic. New possibilities for data acquisition statistics were provided by AIS. It has been designed to supplement the radar information on ships and Vessel Traffic Service (VTS) centres. The system provides a collection of information on vessel traffic through the AIS shore infrastructure in line with the [MSC 74(69), 1998].

AIS combines recording elements of position, course, speed with transponder and broadcasting ship data. AIS ship data is received by other vessels and shore stations equipped AIS along the coast. The data can be visualized on a ship-borne Electronic Navigation Charts (ENC) and VTS centres displays to support the radar information.

According to [MSC 74(69), 1998] AIS should improve the safety of navigation by assisting in the efficient navigation of ships, protection of the environment and operation of Vessel Traffic Services (VTS), through satisfying functional requirements. Navigation safety will depend on reliable information transmitted continuously. Therefore, AIS provides dynamic information refresh in accordance with the time interval from 2 seconds to 3 minutes. The interval depends on the vessel speed, vessel type and course changes. Static and voyage related data are transmitted every 6 minutes, or when the information got altered or on demand. Safety related data is transmitted according to requirements.

The system uses 2 types of devices: Class A (fundamental) and Class B for non-Safety of Life At Sea (SOLAS) vessels.

The time interval for the class A receiver is shown below:

Table 2. 1. Class A shipborne mobile equipment reporting intervals [ITU-R.M.1371, 2013]

illustration not visible in this excerpt

Self-organization of system is done by an algorithm of Time Division Multiple Access (TDMA) providing an update of information, depending on the speed and change course. TDMA algorithm is capable of transmitting signals through over 1000 ships at the same time. For more information on TDMA applied in AIS - [Parkinson A., et al, 2003].

2.3. Information transmitted by AIS used in the dissertation

AIS data were divided into static, semi-static and dynamic information.

a. Static Information: ID Number: MMSI, IMO number, call sign, ship dimensions;
b. Semi-static information: port of destination, type of hazardous cargo, draught;
c. Dynamic information: time stamp, speed over ground, rate of turn, course over ground, geographic position [Aarsaether K.G, Moan T, 2009].

Information used to research the integrity of AIS data is labelled No. 1, 2, 3 - Position Report, and No. 5

- Static & Voyage Related Data. Information used to research the availability of the AIS information channel is message no 4 - Base Station Report.

illustration not visible in this excerpt

Table 2. 2. Message No. 1, 2, 3: Position Report [ITU-R.M.1371, 2013]

illustration not visible in this excerpt

Table 2. 3. Message No. 5: Ship static and voyage related data [ITU-R.M.1371, 2013]

illustration not visible in this excerpt

Table 2. 4. Identifiers to be used by ships to report their type [ITU-R.M.1371, 2013]

illustration not visible in this excerpt

Table 2. 5. Message No. 4: Base station report [ITU-R.M.1371,2013]

illustration not visible in this excerpt

AIS provides vessels and the relevant administrative departments access to information from ships with the required frequency and accuracy. Data transmission should be done with the assumption of the minimum number of staff and a high level of system availability. Dynamic information was used in the calculations performed by Multi-agents Decision Support System (MADSS). The system operates on the basis of received AIS data. It calculates the ship motion parameters. In addition, it generates new motion parameters of the own ship (course, speed) according to designated Closest Point of Approach (CPA). Calculation accuracy of ship motion parameters were shown in [Banachowicz A, Wołejsza P, 2008]. On the other hand AIS availability problem, presented in [Felski A., Jaskólski K., 2011] will be important. The problem of the AIS availability was presented inter alia, by [Hori A., et al., 2009]. Availability of system with the horizontal range is regulated by conditions of radio waves propagation in the VHF band. AIS availability will be reduced under the situation as follows: in case of interference objects present between own vessel and another vessel, in case of multi path and in case of radio wave collision occur in the same slot. The performance of reception greatly depends on the antenna height and horizontal range calculated as follows:

illustration not visible in this excerpt

h1 - antenna height of the vessel transmitting AIS signals; h2 - antenna height of the vessel broadcasting AIS signals.

Adapting the research model to a given area requires knowledge of the system structure in the Gulf of Gdansk. Knowledge of the structure will facilitate the development of a research model of AIS information integrity, data channel availability and the model adaptation to the area where t data recording is conducted.

2.4. AIS network structure of the Gulf of Gdansk

Polish national AIS network was set-up within 2002-2006 in response to the requirements of Copenhagen Declaration signed by Baltic countries in September 2001, which was followed by EC decision expressed in Directive EU 2002/59/EC.

Due to the range of research, the description of the Polish AIS network structure was limited to area of the Gulf of Gdansk. At present, the network of AIS in the Gulf of Gdansk consists of three base stations connected to the regional server, which is connected to the national network server HELCOM. List of Polish base stations in the Gulf of Gdansk is presented in the table below.

Table 2. 6. AIS base stations in the Gulf of Gdansk [Drozd W., Dziewicki M., 2006]

illustration not visible in this excerpt

The base stations are designed to broadcast signals in the VHF band with coverage area of A1 zone. Information derived from AIS base stations can be controlled by Maritime Safety Centre (MSC) in Gdynia, where in real-time a database of AIS information is created. Data from this system is being distributed only to authorized institutions such as Maritime Administration, Polish Boarder Guard, Navy or SAR centre. Wide Area Network (WAN) connects the base stations with regional server located in Gdynia Maritime Office. Regional Server collects information from 3 base stations in the Gulf of Gdansk and is further connected to the server in MSC Gdynia where the national network server of AIS-PL and the database server information were installed.

AIS-PL data stream is down sampled in international server before being sent to the HELCOM server (Copenhagen, Denmark). There data from Poland is integrated with other AIS streams from Germany, Denmark, Russia, Lithuania, Latvia, Estonia, Finland, Sweden and Norway. Finally, the data stream is delivered back to all HECOM members and Norway. Transmission of AIS information in the Gulf of Gdansk is presented by figure 2.1 [Drozd W., et al., 2006].

illustration not visible in this excerpt

Fig. 2. 1. Transmission of AIS information in the Gulf of Gdansk [Drozd W., et al., 2006]

AIS-PL and HELCOM data packets are stored in database server for traffic management, statistics, track plotting, post processing studies or evidence proceedings. Stored data is periodically backed up onto a tape storage system. Additionally - data transmission from/to HELCOM server is encrypted for security reasons by Secure Socket Layer (SSL) protocol [Drozd W., Dziewicki M., 2006].

The main task of national server is to support regional servers. National server is located in Gdynia VTS Centre. The regional server of Gulf of Gdansk is also located there.

Receiving and provision principles of data were determined by the Copenhagen Declaration signatories. The basic principle of data transfer is limited access just for authorized institutions. Each participant can filter the data in a way approved by him, and could host AIS data from its AIS national network via Internet and the data format must be compatible with IEC 61993-2. The period of data update cannot exceed 6 minutes. Each participant of the Declaration is responsible for the proper use of data by competent authorized institutions. AIS data provider must take appropriate preventive action prior to redistribution to third parties. Participants declare their free will to maintain the correct transmission of AIS data. However, the data transmitting party or receiving party does not provide any warranty in relation to AIS data availability, accuracy and integrity [Dziewicki M., Ledóchowski M., 2005].

In 2010, the international experiment „CAP DEMO” was conducted, which emphasized the usefulness of the MarSSIES System - Maritime Safety and Security Information Exchange System and AIS in the fight against terrorism and illegal activity at sea. Task was carried out by combining the Polish and international system of supervision (Singapore OASIS, Spanish COVAM, Swedish SUBCAS and Finnish MEVAT). The subject of “CAP DEMO” was to monitor, detect and neutralize the threat caused by the appearance of a ship suspected of terrorist activities. The threat was simulated by warship ORP "ARCTOWSKI" which at the time of experiment cruised with false identification data. Identification data belonged to a ship cruising in area of Singapore. System MarSSIES automatically detected the error, generated alarm and initiated action with the use of Polish Navy, Maritime Regional Unit of the Border Guard Republic of Poland (MOSG), Maritime Offices and Special Forces [Dura M., 2011].

The experiment proved that AIS can help in the ongoing connecting process of marine surveillance systems in Europe. This can be a software platform for information distribution between the operational services cooperating in the field of maritime safety.

According to SOLAS Convention, AIS is a fully operational system. AIS was designed as a terrestrial system based on the use Very High Frequency (VHF) radio frequencies (161,975, 162,025MHz) to transmit data. The reception of VHF signals by satellites is a proven technology, with many commercial applications. AIS Satellite use VHF maritime mobile band and Ultra High Frequency UHF radio reception via satellites. Operating in use 137.000 - 150.500 MHz and 400.075 - 400.125 MHz frequency ranges, satellite reception of these signals can be cost-effective tracking, monitoring and messaging capability. Details concerning AIS Satellite were contained in [Carson-Jackson J., 2012]. However, there are problems related to the system functioning with limited availability and integrity of system information. These problems relate to the antenna height, signal interference, multipath affecting the reliability of the system, which have been generally investigated.

2.5. AIS malfunctions

Since the end of AIS implementation, ships covered by SOLAS Convention were obliged to have on-board AIS.

Initially, many watch-keeping officers were under the impression that AIS was a "fit and forget” equipment. It was probably caused by lack of proper training or misinterpretation of the IMO performance standards stating: [MSC 74(69), 1998] „ The AIS should be capable of providing information automatically and continuously to a competent authority and other ships, without ship ’ s personnel involvement”.

During the AIS implementation manufacturers faced the problem of digital interface incompatibility with most radar systems and Electronic Chart Display Information System (ECDIS).. Display of AIS information on radar indicators was associated with high costs [Norris A., 2007]. Much cheaper way was the presentation of AIS information on monochrome indicators called Minimum Keyboard Display (MKD) with typical size of screen 120x90 mm. In this case, it is justified to formulate questions: Can the data presentation on MKD indicators effectively illustrate the individual movement in area of heavy marine traffic, on the basis of the CPA and TCPA?

Imperfection of AIS functioning is the lack of information for the user on how to correctly set up the device for operation. This problem results from the lack of proper training in the field of device operation. Manuals supplied with the purchased equipment were incorrect and as a result incorrect static data was being entered by users. Voyage related data: ports of destination, ETA, navigation status were not entered. In addition, the use of safety related data were not properly understood by users.

Incompatibility of GPS units with AIS also became problematic. If the receiver is not in the operating range of the base station, then the lack of GPS signal may be the cause of no transmission of AIS signals. Additionally, if the GPS receiver operates in a system other than WGS-84 then it sends out position information with an error. This issue should be considered mainly by staff installing the AIS system on board.

Proper positioning of the ship's AIS antenna needs to be taken into account during the AIS installation. Poorly sited AIS antennas cause interference to AIS from other systems, which results in data corruption and limits signals transmission. AIS antenna should be mounted in accordance with IMO AIS Guidelines S/N Circ 227.

Interference occurs also when an incorrect antenna cable is used. In such case number and quality of transmitted signals is limited. It results in AIS information not being received system planned interval. AIS test sets, which could have identified these problems on installation, were not available for earlier installations [Norris A., 2007].

There were problems with heading information. Not all ships had a heading indicator with a digital interface. Bridge staff were often not informed on how these should be checked and adjusted, resulting in transmitted headings having possible offsets of up to 180o [Norris A., 2007].

In many instances bridge staff remains unaware of the anomalies for a substantial time. They had not been informed of the necessity for regular checks on transmitted data, because port and coastal authorities did not necessarily have AIS facilities, defects were not being routinely communicated to errant ships.

The issue of the AIS functioning leads to reflection on performing detailed studies of information integrity.

In 2007, the results of the research were presented by Liverpool John Moores University [HaratiMokhtari et al., 2007a]. Limitations of AIS information integrity were mainly related to human errors and input failure of static data. The errors were categorized in the following way:

a. Errors committed by the operators, users;
b. Installation Failures;
c. Design failures;
d. Training and management levels;
e. Regulatory failures.

In accordance with [Harati-Mokhtari et al, 2007b] 80-85% of all recorded maritime accidents resulted from human errors. Contribution of human error to maritime accidents has increased over the period of 1991-2001. Most accidents are the result of senseless and avoidable human errors. Human errors depend on internal factors related to the personal characteristics of operators such as: qualifications, acquired skills, work experience. In addition, there are also external factors such as: equipment design, device installation, the complexity of the task, the work environment, organizational factors, operating procedures. Thus the proper balance between the device operator and the complexity of tasks reduces the likelihood of human error [Harati-Mokhtari et al, 2007b].

Facts about the system imperfections lead to conclusion that static information errors result from operators’ mistakes and unit installation errors. Sensor errors and factors responsible for limiting the system availability have reduced the AIS integrity. Interference signals, radio propagation conditions have an impact on coverage area. In this context, the question arises: is it reasonable to introduce AIS as an element affecting the increase in safety and efficiency of navigation, when the AIS may have low information integrity?

On the basis of pilot studies conducted by author, it is stated that the level of confidence in the AIS information does not only depend on human error. Information inadequacy is limited by message errors originating form sensors. Thus, true heading inadequacy is 16,54% and rate of turn inadequacy is 22,20%.

It was established that, research unfitness of AIS information is a complex process and requires development of research method estimating the completeness and integrity coefficients of 3 types of information - dynamic, static and voyage related data. This information is contained in message No. 1, 2,

3 (Position Report) and message No. 5 (Static and voyage related data). Another factor determining the level of confidence in the availability of AIS binary data channel based on message No. 4: Base Station Report. Analysis of the AIS availability is presented in [Felski A., Jaskólski K., 2011a]. In accordance with accepted hypotheses, stochastic model based on the research outcomes, is presented in this dissertation.

3. REVIEW OF AVAILABILITY AND INTEGRITY METHODS AND TOOLS FOR AIS INFORMATION RESEARCH

3.1. Methods and tools for research of AIS information integrity

There is much research on AIS information integrity and availability. The problem was analysed, for example by [Harati-Mokhtari A., et al., 2007a], [Bailey N., 2005], [Drozd W., et al., 2007], [HaratiMokhtari A. et al., 2007b], [Hori A., et al., 2009]. Outcomes of the research methodology of this topic were also presented in [Felski A., Jaskólski K., 2008]. Researches on the AIS coverage area and the system availability were shown in [Hori A. et al., 2009], [Lapinski S., Isenor W., 2011], [Hammond T, Peters D., 2012], [Jaskólski K., 2012].

3.1.1. Application of statistic methods to research AIS service information

The method of statistical analysis of AIS information unfitness was presented by [HaratiǦMokhtari

A. et al., 2007]. The authors defined the types of mistakes and errors made by AIS system users. Statistical analysis of AIS information quality is based on the modification of reliability system model referred to as “Swiss Cheese Model”.

illustration not visible in this excerpt

Fig. 3. 1. „Swiss Cheese” model used in AIS information quality research [Harati-Mokhtari A. et al., 2007]

It was estimated, that at least 80% of all recorded maritime accidents resulted from human errors.. Therefore, the contribution of human error as a main factor leading to the increased number of maritime accidents was studied over a ten-year period of 1991-2001 [Harati-Mokhtari A. et al, 2007]. Reliability studies of AIS information were conducted by Liverpool John Moores University in three locations:

a. VTS Liverpool - form September to October 2005, where information about 94 vessels was recorded;
b. AISLive Company of Lloyds Register-Fairplay Ltd - from the 1st of March to the 17th of March 2005, where 400059 reports were recorded, of which 30946 reports were assessed;
c. Service AISweb of Dolphin Maritime Software Ltd - form 23.11.2005 to 02.05.2006.

The criterion of reliability assessment of AIS information was presented in Table 3.1. Information status was designated under the condition defining the unfitness state. The method defined the reliability criteria only to limited extent, with minimum level of system technical specification used.

Table 3. 1. Criterion of the assessment of AIS information reliability [Harati-Mokhtari et al, 2007]

illustration not visible in this excerpt

3.1.2. Statistical analysis of AIS information in accordance with N.Bailey theory

At the same time i.e. in the year 2005, similar research of AIS information has been conducted in Dover Strait which is considered to be a narrow and high traffic intensity channel area [Bailey N, 2005]. The studies on reliability and completeness of AIS information have been conducted by the comparison of two sources of information. Static information and voyage related data have been compared with verbal information. Verbal information was reported on Channel 16 (156,8 MHz) to VTS operators upon the entrance to and the departure from VTS area, respectively. It was assumed that information transmitted by radiotelephone was reliable, whereas AIS information might contain errors. Research details were divided into categories depending on transmitted information. Percentage distribution of AIS static information was presented in Figure 3.2 [Bailey N, 2005].

[...]