Fig 7 - uploaded by Francisco Lazaro
Content may be subject to copyright.
LEO satellite ground track and selected positions where the throughput is evaluated. The LEO satellite is assumed to have a circular orbit and an altitude of about 560 km. Assuming that ships with at least 0 degrees of elevation angle can be received at the satellite, the corresponding field of view of the satellite is about 2500 km in radius, as previously mentioned.
Source publication
Satellite-aided Automatic Identification System (AIS) has drawn the attention from the maritime community for worldwide vessel tracking and monitoring. The satellite reception of AIS packets is viable, although the standard has been designed for inter-vessel communication and vessel collision avoidance. A model and an extensive analysis of the Medi...
Context in source publication
Context 1
... the assumption considered so far of having ships uniformly spread over a geographical area has helped in verifying the validity of the model, two questions arise: What is the the range of channel loads G seen at a LEO satellite? Depending on the answer, in fact, different countermeasures and different advanced signal processing techniques can be exploited. The second question is whether the model is accurate enough also when realistic ship distributions come into play, which may create ties across overlapping clusters of ships that were disregarded in the developed analytical framework. For this, in the next subsections we first characterize the load, and then study how the throughput of the system behaves in such conditions. In this subsection we focus on the simulation of the channel load G seen from a LEO satellite with the realistic ship distribution (Fig. 4). The results of the simulations, reported in Fig. 6, are provided for the area delimited by the north part of the Atlantic Ocean and Europe. Each simulated position of the satellite is evaluated for 15 SOTDMA frames. Two different simulation instances per satellite point are performed for increasing the confidence of the results. The areas with the highest density of AIS packets are around the Portuguese and Spanish coasts, the Strait of Gibraltar and in the Middle East Area, close to the Suez canal and the Red Sea (see Fig. 6). In these areas the channel load G can exceed 6 packet per slot. On the other hand, in the central part of the Atlantic Ocean, when we are close to Greenland and in the area of Finland/Russia, the channel load G can be lower than 1 packet per slot. Since the ship density data are not very precise in the Baltic Sea area and in the Mediterranean Sea (as it is discussed in Fig. 4), we can expect higher channel loads in a real scenario than the one simulated (which is around 3 packets per slot in the former and around 4 packets per slot in the latter). The answer to the question raised in the previous subsection is thus that the channel load ranges G from a LEO satellite perspective spans from less than 1 packet per slot to more than 6 packets per slot, and confirms the validity of the choice for the evaluation done in the previous subsection. It is now important to verify if the model developed can also precisely predict the throughput of an SOTDMA system where the ships are distributed realistically on the Seas surface and the assumption of independent cluster can be inaccurate. This is the objective of the next subsection. In this subsection we focus on the simulation of the throughput for a realistic ship distribution. We define a possible LEO satellite ground track (see Fig. 7) and we select on this ground track six satellite positions where the throughput will be evaluated. The results are compared with the uniform ship distribution and with a SA system with same channel load. The six marked points are the satellite positions where the throughput vs. channel load is evaluated. For every selected satellite position, the results are averaged over 10 different simulations of 15 SOTDMA frames each. Also in this case, we assume realistic ship distribution (see Fig. 4). In Fig. 8 the results are reported together with the performance achieved under a uniform ship distribution and the equivalent SA system. The first observation is that the SA model is very close also to the results with realistic ship distribution for all the satellite positions. Furthermore, for channel loads below 2 packets per slot (points D, B and F) SOTDMA is able to slightly outperform SA in terms of throughput, while for channel loads above 2 packets per slot (points E, A and C) the opposite happens. This results is precisely predicted by the model developed in this paper (see figures 2 and 3). VI. R ANDOM A CCESS AND S ATELLITE AIS The discussion carried out in Sections IV and V has highlighted how the behaviour of a system formed by several clusters of ships resorting to SOTDMA can be very well modeled at a satellite by a random access scheme where all transmissions take place in an uncoordinated fashion. Such a result is particularly handy, as it allows to accurately predict in a vast range of conditions the performance of a complex system without the need to characterize all the details of a distributed reservation-based medium access control. On the other hand, it also hints at the possibility of using a simpler solution for receiving messages at the satellite with comparable or higher efficiency. This comes as no surprise, considering the large number of vessels out of each other’s reception range that may fall within the footprint of the same satellite without being able to properly coordinate. In this perspective, even though the SOTDMA-based AIS protocol can be regarded nowadays as a well-established and enduring standard with a very high level of penetration for ship-to-ship and ship-to-shore communications, the availability (and likely expansion) of dedicated VHF channels for transmission of maritime traffic over satellite [8] paves the road for the usage of alternative and more efficient medium access solutions. Starting from these remarks, the present paper aims at trigger- ing a discussion on the definition of novel and viable random access schemes tailored to AIS reception at the satellite. A first remark along this direction immediately follows from the results presented in the previous sections. It is in fact clear that, even without any modification to the transmission intervals mandated by the AIS standard, having ships transmit following Slotted Aloha would improve performance. The benefit would in this case be twofold, as a higher throughput would be granted when a large number of vessels fall within the satellite footprint while avoiding the latency induced by the learning phase of SOTDMA. Secondly, the simulation study reported in Fig. 6 prompts how the overall traffic level perceived at the satellite is particularly high (e.g., more than 5 pk/slot) in many geographical regions that are of practical relevance. This issue stems from the transmission rates foreseen by AIS, originally designed to allow real-time tracking of ships within each other’s coverage range and tailored for a small number of users contending for a TDMA frame rather than for the very large population seen from a LEO orbit. Such conditions significantly deteriorate the capability of successful message retrieval, and represent one of the key challenges to be faced in satellite-AIS systems. From this standpoint, even though a lot of effort has been devoted to advanced signal processing solutions capable of decoding even in the presence of collisions [9], [5], the amount of information that can be collected is intrinsically limited by the channel overload. Conversely, if we tackle the problem from a link layer perspective, the parallel between SOTDMA and SA allows to straightforwardly characterize the tradeoff between the transmission intervals employed by a ship and the actual frequency with which it can be seen at the satellite. Let us in fact consider a topology where n vessels are spread within the footprint of interest. As discussed in Section II, we can divide AIS transmitters into four groups according to their speed, such that the η i n ships in set C i send on average φ i = α i φ ∗ messages per slot, with i η i = 1 and φ ∗ a reference packet generation frequency. If we model the uplink as a random access channel, and thus take advantage of the independent transmission pattern assumption among users, the average load seen at the satellite can be readily estimated as G = κnφ ∗ , where κ = i α i η i . Resorting to the Poisson traffic approximation discussed in Section IV, the frequency ρ i with which a message of a ship in C i is received uncollided evaluates ...
Citations
... In an initial review of our findings, the correlation between highly congested areas and dark vessels was noted [14]. This phenomenon can be attributed to a possible high number of packet-collisions in the area [72,73], where, due to high traffic, some AIS messages are not being received. In order to counter that effect and to differentiate between low AIS coverage due to packet collision and intentional switch-off of AIS transponders, we performed an analysis described in this section, which is based on identifying areas (e.g., grid cells) where vessels are in close distance, thus increasing the risk of losing AIS packets due to congestion. ...
Density maps support a bird’s eye view of vessel traffic, through providing an overview of vessel behavior, either at a regional or global scale in a given timeframe. However, any inaccuracies in the underlying data, due to sensor noise or other factors, evidently lead to erroneous interpretations and misleading visualizations. In this work, we propose a novel algorithmic framework for generating highly accurate density maps of shipping activities, from incomplete data collected by the Automatic Identification System (AIS). The complete framework involves a number of computational steps for (1) cleaning and filtering AIS data, (2) improving the quality of the input dataset (through trajectory reconstruction and satellite image analysis) and (3) computing and visualizing the subsequent vessel traffic as density maps. The framework describes an end-to-end implementation pipeline for a real world system, capable of addressing several of the underlying issues of AIS datasets. Real-world data are used to demonstrate the effectiveness of our framework. These experiments show that our trajectory reconstruction method results in significant improvements up to 15% and 26% for temporal gaps of 3–6 and 6–24 h, respectively, in comparison to the baseline methodology. Additionally, a use case in European waters highlights our capability of detecting “dark vessels”, i.e., vessel positions not present in the AIS data.
... For shipborne AIS, transmitted information may include dynamic information, providing live information about the vessel, such as a ship's geographical coordinates, course over ground (COG), speed over ground (SOG), etc., as well as static information that is relatively permanent to a particular vessel, including ship's name, Maritime Mobile Service Identity (MMSI), dimensions, etc., along with voyage-related information specific to each voyage, such as the ship's navigational status, destination, estimated time of arrival (ETA), etc., [1,2,5]. Providing this dynamic, static, voyage-related, and other additional information, AIS has improved both the safety of navigation on water as well as an enhanced level of surveillance from land [6], but it still faces some difficulties. ...
Traditional methods of marine navigation are undergoing a revolution brought about by the almost universal adoption of the Automatic Identification System (AIS). AIS exchanges a wealth of navigational information among vessels and between ships to shore through Very High Frequency (VHF). With AIS data integrated into the Electronic Chart Display and Information System (ECDIS), the identification and navigational information of surrounding vessels as well as aids to navigation can be reflected on the electronic charts in real time, despite some problems such as the low AIS carriage rate on small vessels where it is not mandatory and the high cost of ECDIS preventing such vessels from installing it. In this paper, we introduce BlueNavi, a lower cost but sustainable maritime information providing platform built with microservices architecture allowing flexible on-demand scalability and cross-platform adaptability. Applications served by BlueNavi can provide users with data either stored in a remote data center through the internet or received locally by devices connected to the station without the need for the internet. From our land test, we show that users with only an internet connection but without any AIS equipment can also obtain live AIS data collected by other stations. Conversely, with access to the internet, BlueNavi can also send data back to the land stations, enabling other ships to identify non-AIS ships as well. Through the live-ship test, we demonstrate that BlueNavi works well offline in cooperation with shipborne AIS equipment. We also look at some possible application scenarios for BlueNavi with other data sources and means of communication other than AIS and VHF that can be expanded to the platform. BlueNavi will enable inexpensive ship identification for small vessels and provide an extension of functionality to ECDIS for large ships.
... Among these sources lies the Automatic Identification System (AIS), which automatically collects messages from vessels around the world, at a high frequency. AIS data basically consist in GPS-like data, together with the instantaneous speed and heading, and some vessel specific static information (Clazzer et al., 2014). An example of such data is presented in Fig. 1 (left), considering 6 months of AIS data of vessels steaming in the Ushant traffic separation scheme in Brittany, west of France (Fablet et al., 2017). ...
In the context of the surveillance of the maritime traffic, a major challenge is the automatic identification of traffic flows from a set of observed trajectories, in order to derive good management measures or to detect abnormal or illegal behaviours for example.
In this paper, we propose a new modelling framework to cluster sequences of a large amount of trajectories recorded at potentially irregular frequencies. The model is specified within a continuous time framework, being robust to irregular sampling in records and accounting for possible heterogeneous movement patterns within a single trajectory. It partitions a trajectory into sub-trajectories, or movement modes, allowing a clustering of both individuals’ movement patterns and trajectories. The clustering is performed using non parametric Bayesian methods, namely the hierarchical Dirichlet process, and considers a stochastic variational inference to estimate the model’s parameters, hence providing a scalable method in an easy-to-distribute framework. Performance is assessed on both simulated data and on our motivational large trajectory dataset from the automatic identification system, used to monitor the world maritime traffic: the clusters represent significant, atomic motion-patterns, making the model informative for stakeholders.
... As highlighted in [11], KPIs that cannot be expressed as closed form functions of system parameters require extensive simulations for the performance evaluation. Even though authors of [24,25] have obtained closed form expressions of probability of collision, the expressions are subject to strict assumptions that restrict the scope of their application. We predict, using a CSS approach, the variation of aforementioned KPIs using functional complexity. ...
... As for the modulation, a Gaussian Minimum Shift Keying (GMSK) modulation is used, with a modulation index h = 1 2 and a product B × T = 0.3 or 0.5, where B is the −3 dB cut-off frequency of the Gaussian filter and T the bit duration (Berder et al., 2005). No interleaving nor Forward Error Correction (FEC) is used (Clazzer, Munari, et al., 2014). In addition, the brevity of ship transmissions cause inter channel interference (ICI) and decrease the rate of successful transmission. ...
... This could be useful in the case in which heavy collision is observed, such as satellite reception. As for satellite reception, (Clazzer, Munari, et al., 2014) demonstrate that SOTDMA can be seen as a slotted random access protocol at the satellite. This simplifies in a noteworthy way the analysis of the protocol performance, as an optimisation on the rate of transmission of AIS packets generation can be done to have a better success rate and then maximise the vessel tracking frequency. ...
At sea, various systems enable vessels to be aware of their environment and on the coast, those systems, such as radar, provide a picture of the maritime traffic to the coastal states. One of those systems, the Automatic Identification System (AIS) is used for security purposes (anti-collision) and as a tool for on-shore bodies as a control and surveillance and decision-support tool.An assessment of AIS based on data quality dimensions is proposed, in which integrity is highlighted as the most important of data quality dimensions. As the structure of AIS data is complex, a list of integrity items have been established, their purpose being to assess the consistency of the data within the data fields with the technical specifications of the system and the consistency of the data fields within themselves in a message and between the different messages. In addition, the use of additional data (such as fleet registers) provides additional information to assess the truthfulness and the genuineness of an AIS message and its sender.The system is weekly secured and bad quality data have been demonstrated, such as errors in the messages, data falsification or data spoofing, exemplified in concrete cases such as identity theft or vessel voluntary disappearances. In addition to message assessment, a set of threats have been identified, and an assessment of the associated risks is proposed, allowing a better comprehension of the maritime situation and the establishment of links between the vulnerabilities caused by the weaknesses of the system and the maritime risks related to the safety and security of maritime navigation.
... The technical characteristics behind the AIS system are published in Recommendation ITU-R M.1371-5 [9]. In [10], Clazzer et al. analytically model the AIS Self-Organized Time Division Multiple Access (SOTDMA) traffic pattern at the satellite and investigate the realistic behavior of SOTDMA via simulations. Shelmerdine [11] demonstrates a procedure for the processing, analysing, and visualisation of AIS data with example outputs and their potential uses. ...
In this paper, Automatic Identification System (AIS) data collected from space is used to demonstrate how the data can support search and rescue (SAR) operations in remote waters. The data was recorded by the Norwegian polar orbiting satellite AISSat-1. This is a case study discussing the Ortelius incident in Svalbard in early June 2016. The tourist vessel flying the flag of Cyprus experienced engine failure in a remote part of the Arctic Archipelago. The passengers and crew were not harmed. There were no Norwegian Coast Guard vessels in the vicinity. The Governor of Svalbard had to deploy her vessel Polarsyssel to assist the Ortelius. The paper shows that satellite-based AIS enables SAR coordination centers to swiftly determine the identity and precise location of vessels in the vicinity of the troubled ship. This knowledge makes it easier to coordinate SAR operations.
... employed in AIS and originally devised for communications among neighbouring ships. When used to send messages to a satellite, however, this approach has been shown to incur significant throughput degradation in areas with high traffic density [4]. Due to the very large footprint of a low earth orbit (LEO) satellite, in fact, the benefits of distributed medium coordination among vessels in reciprocal reception range are lost, leading to packet collisions at the receiver and to performance similar to a slotted ALOHA (SA) scheme [5]. ...
The paper focuses on the satellite uplink for the new VHF data exchange system (VDES), which aims at providing a worldwide messaging service for vessels. In this context, we investigate the use of some recently proposed asynchronous schemes for the narrowband random access mode of the standard. Remarkable improvements over the basic VDES self-organised TDMA return link access are shown in terms of both spectral efficiency and system coverage when considering realistic ships distributions. The study offers relevant insights for the ongoing discussions on the MAC layer design for upcoming versions of the VDES standard.
... Another drawback of the maritime VHF radio is its lack of support of data services. Maritime satellite communication systems [1,2], such as the Inmarsat-F system and Fleet-Broadband maritime data service, are suitable for ocean sea ship communications. However, the satellite communication system is relatively expensive, due to the high cost of the terminal equipment and high maintenance and upgrade costs. ...
... In (iii), ( , ) is the Euclidean distance metric; is the average distance associated with the reflections. Proposition 2. The Shannon maximum entropy ( , ), which is described by formula (2) and satisfies condition C1, can be expressed as ...
Multipath effect in vessel communication is caused by a combination of reflections from the sea surface and vessels. This paper proposes employing stochastic ray method to analyze maritime multipath propagation properties. The paper begins by modeling maritime propagation environment of radio waves as random lattice grid, by utilizing maximum entropy principle to calculate the probability of stochastic ray undergoing
k
time(s) reflection(s), and by using stochastic process to produce the basic random variables. Then, the paper constructs the multipath channel characteristic parameters, including amplitude gain, time delay, and impulse response, based on the basic random variables. Finally, the paper carries out a digital simulation in two-dimensional specific fishery fleet model environment. The statistical properties of parameters, including amplitude response, probability delay distribution, and power delay profiles, are obtained. Using these parameters, the paper calculates the root-mean-squared (rms) delay spread value with the amount of 9.64
μ
s. It is a good reference for the research of maritime wireless transmission rate of the vessels. It contributes to a better understanding of the causes and effects of multipath effect in vessel communication.
... In order to distribute up-to-date information to neighbouring vessels, messages are sent more often as the ship speed increases, resulting in 4 possible frequencies of packet transmission. For more details on SOTDMA, refer to [1], [7]. this issue affects dramatically the tracking performance of a satellite-aided AIS system, all the more so considering the steadily increasing traffic generated by other maritime communication services being allocated to VHF band [8]. ...
... On the other hand, a unified framework capable of accounting for topological, MAC and physical (PHY) layer issues and flexible enough to support different system configurations to gather a clearer understanding of the main performance drivers is still missing. A first step in this direction was taken by the work in [7], showing that the MAC protocol of AIS as perceived at a satellite can be well approximated resorting to random access such as slotted ALOHA (SA). This result is particularly interesting, as it allows to abstract the non-trivial details of SOTDMA and characterise the performance at the MAC layer through a single channel traffic coefficient. ...
... A remarkable outcome will be the assessment of the satellite reception capabilities of AIS messages with advanced receivers that may apply multi-user detection (MUD) techniques as, for example, successive interference cancellation (SIC). On the other hand, the MAC layer will exploit the model of SOTDMA as SA derived in [7]. ...
In this paper, we discuss the impact of the capture effect and multiuser detection on the reception of automatic identification system messages at the satellite. Resorting to simple yet beneficial tools such as the multi-packet reception matrix and the random access approximation of the medium access layer, an insightful model is presented. A comparison with the destructive collision model is also carried out and performance evaluation with vessel distribution and vessel speed distribution gathered from real data are derived. The more realistic physical layer improves the results accuracy and provides tradeoffs and future research directions.
... An interesting characterization of the problem was given in [9], where the authors showed by means of a simple analytical framework how incoming AIS traffic at a Low Earth Orbit (LEO) satellite can be very accurately modeled considering a slotted Aloha access scheme that disregards any form of coordination among vessels. Taking the lead from this, and relying on experimental data for ships distribution, it was possible to prove how the average load to be expected in regions of interest such as the Mediterranean see or the western coasts of Europe can easily be very high (e.g, larger than 5 pk/slot). ...
... Within this paper, we study this tradeoff in greater detail, deriving some insights of interest both from a research and a practical viewpoint. Extending the work in [9], we develop an analytical model to characterize relevant metrics such as throughput and detection probability. To achieve this we work at packet level, considering collisions among messages as destructive and abstracting physical layer details. ...
... An exact characterization of the traffic pattern generated as per SOTDMA by ship clusters falling within the footprint of the receiver and not coordinating among each other is in general not trivial. On the other hand, [9] showed that a very good approximation can be obtained assuming all vessels to simply generate messages according to a Poisson process of aggregate intensity G and accessing the medium as soon as data units are available for transmission. Leveraging this result, we model the flying platform as the receiver in a well-known Slotted ALOHA (SA) protocol. ...
In the recent past, an increasing interest has been devoted to the possibility of receiving Automatic Identification System (AIS) messages via Low Earth Orbit (LEO) satellites. While the principle has been demonstrated to be a viable option for monitoring vessel traffic over oceans and vaste land areas, the achievable performance from a communications viewpoint is far from optimal. Recently, it was shown how AIS traffic seen at a satellite can be very accurately modeled resorting to simple random access schemes. Leveraging this result, in this work we propose a simple yet flexible analytical framework capable of predicting channel load and overall reception performance taking into account the spatial distribution of vessels as well as their traffic generation pattern. Feeding the model with ship speed and location data derived from experimental settings, we discuss the achievable efficiency for a typical LEO-satellite detecting AIS packets. Moreover, the impact of the receiver footprint on ground on the overall decoding performance is investigated, deriving some interesting insights on the benefits that could stem resorting to narrower-beam systems. In this direction, we discuss two cases: the usage of a LEO satellite with a directional antenna soon to be launched for AIS monitoring, and the possibility of using airliner for receiving vessel-generated traffic.
