The seas are becoming increasingly congested, with the IFT Transport Outlook 2019 predicting that demand for maritime trade could triple within the next 30 years. A large number of ships raises the chances of the collision. Although the number of serious collisions has fallen within the past decade with 132 lost in 2009, compared with 46 in 2019, the rise in traffic is causing safety concerns. This is because even a minor collision can be enough to disrupt operations, posing risks to crew and the marine environment. In the 23,073 marine casualties and incidents that happened from 2011 to 2018, the total number of ships involved was 25,614. General cargo ships were the main category involved in a marine casualty or incident (43.8%), followed by passenger ships (23.7%).
Collisions not only result in loss of life or damage to property but also huge environmental impacts. The fisheries and marine wildlife could be impacted for years to come. The environmental impact depends on many factors such as the size of the spill, its chemical makeup, and where the chemicals from the oil will spread.
When a vessel is navigating at sea, the movement and identity of other vessels in the vicinity are critical for navigators to make decisions to avoid a collision. Captains need to be able to maneuver their ships within feet in the worst of conditions and to be able to navigate “blind” when there is no visibility at night or due to bad weather.
Vehicle collision is also a growing problem for unmanned surface vehicles (USVs), also known as unmanned surface vessels, that have attracted significant attention for their potential applications to performing time‐consuming and/or dangerous missions such as patrol, surveillance and reconnaissance, environmental monitoring, and inspection of marine structures.
Causes of Marine accidents and Collisions
There have been many reasons for ship collisions. The collisions are also further enhanced due to human errors, equipment failures or external environmental factors.
It has been estimated that human error is responsible for anywhere between 75% and 96% of all marine accidents. Some collisions are reportedly caused by crew members becoming too consumed with onboard instruments, meaning that they fail to look at what is happening on the seas outside.
Others have apportioned some of the blame on the reduced crew numbers since the 2008 global financial crisis forced companies to operate in tighter margins, requiring crew members to do more work than they would have previously done. This can lead to fatigue and stress for crew members, reducing performance. Distraction from smartphones has been another given reason, as well as a lack of training and experience on some vessels.
Naval vessels have been also involved in collisions. In 2017, Seven U.S. sailors were killed when the USS Fitzgerald collided with a Philippine-flagged container ship in the middle of the night off the coast of Yokosuuka, Japan, June 17. The collision between Fitzgerald and Crystal was avoidable and resulted from an accumulation of smaller errors over time, ultimately resulting in a lack of adherence to sound navigational practices. Specifically, Fitzgerald’s watch teams disregarded established norms of basic contact management and, more importantly, leadership failed to adhere to well-established protocols put in place to prevent collisions. In addition, the ship’s triad was absent during an evolution where their experience, guidance and example would have greatly benefited the ship.
Ten U.S. sailors were killed when the American destroyer USS John S. McCain, collided with commercial vessel Alnic MC in waters east of Singapore on Aug. 21, according to the Navy. Singapore Strait is one of the world’s busiest seaways leading into the Strait of Malacca where half of the world’s seaborne shipments by tonnage pass through. These vessels are sometimes separated by under a nautical mile, or about 1.8km.
The collision between John S. McCain and Alnic MC was also avoidable and resulted primarily from complacency, over-confidence and lack of procedural compliance. A major contributing factor to the collision was sub-standard level of knowledge regarding the operation of the ship control console. In particular, McCain’s commanding officer disregarded recommendations from his executive officer, navigator and senior watch officer to set sea and anchor watch teams in a timely fashion to ensure the safe and effective operation of the ship. The impact of the tanker ripped a hole in the hull flooding three berthing areas and drowning ten sailors who were unable to escape from the spaces.
Another reason for the nondetection of modern warships is that they are designed with stealthy features such as painted with radar-absorbent paints and having angular features to reflect signals elsewhere and making it appear as a fishing boat to radar. The stealth features also make it more accidental, “As soon as the ship first went to sea, complaints from alarmed merchant ships poured in. Their radar saw “just a fishing boat,” and when the destroyer was in sight, they were forced to urgently scramble to change course,” said a Russian article.
There is also a rising threat of cyber attacks on GPS and other systems. This form of attack involves overpowering the receiver by broadcasting signals that are synchronized with the legitimate signals detected by it, thereby forcing GPS to provide false information. In July, 2017 the US Maritime Administration reported an incident in which at least 20 Russian ships appeared on trackers to be in the same spot 20 miles (32 kilometres) inland, despite being at various positions in the Black Sea. While this initially appeared to be a glitch, experts now suggest that Russia may have been testing a new system for spoofing GPS.
Military experts said there is a high chance the US warship did not have its AIS activated, a common security practice among military ships. Naval ships are exempt from the international requirement that vehicles autonomously and continuously broadcast their position, course and speed. But Mr Bitzinger said the bigger challenge facing the US Navy is one of fatigue and complacency. Both crews did not attempt to contact the merchant ship bearing down on them, sound a warning horn, sound a collision warning or sound general quarters before the impacts.
A navigator should always be on the bridge to evaluate/re-evaluate the collision risk with respect to expected or unexpected ship behaviour resulting from his course and speed control actions. Unexpected ship behaviour can complicate the navigator`s decision-making process in some situations, where an adequate understanding of the situation may not be possible. When on-board systems are making the same decisions for future autonomous vessels, this can further complicate not only those system decisions but also their interactions (i.e. the outcomes) with the decisions made by manned vessels (i.e. human decisions).
International Regulations for Preventing Collisions at Sea (COLREGS)
Ship encounter situations, related to possible near-miss and collision situations, are regulated by the International Regulations for Preventing Collisions at Sea 1972 (COLREGs) in open sea areas. The IMO Convention on the International Regulations for Preventing Collisions at Sea (COLREGS) was adopted in 1972 and came into force in 1977, replacing the 1960 Collision Regulations. But there have been laws governing the seas for hundreds of years.
COLREGS are in place to ensure that all those operating vessels adhere to a necessary safety framework, covering aspects such as the best way to pass oncoming vessels, safe speeds in certain locations, the safest way to overtake, as well as required equipment that includes tracking devices. Regulations apply to all vessels operating on the seas, with special rules for different sizes and classes of ships, from small sailing vessels and fishing boats to large-scale oil tankers.
Furthermore, additional local navigation rules and regulations can be enforced on ships, especially in confined waters and maritime traffic lanes. On the other hand, if the collision risk can be detected relatively far away from a ship encounter situation, then vessels can take appropriate actions to avoid even a close encounter situation. That step can eliminate the possibility of any close ship encounter situation.
Collision avoidance
Collision avoidance system for ships is an autonomous system that examines risk of collision between ships based on the data from navigational equipment and performs appropriate avoidance control to prevent the collision.
An important element for Collision avoidance is Situation Awareness, that is the perception of the elements in the environment within volume of time and apace, the comprehension of their meaning, and the projection of their status in the near future. Situation Awareness may be gained by answering four simple questions; What happened?; Where are we?; What is happening?; What could happen? There are three components to gaining Situation awareness; gathering data, understanding and projecting ahead. In order to maintain and improve Situation awareness, one should repeatedly go through the three components.
For safe USV operation, autonomous navigation technologies, including path planning, guidance and control, obstacle detection, and mapping, are required. In particular, automatic detection of surrounding objects and their motion estimation are key aspects, and reliable and robust performance in a wide variety of environmental conditions are necessitated. To provide situational awareness various perception sensors such as radar, lidar, and cameras have been mounted on the USV platform and automatic ship detection algorithms are applied to the sensor measurements.
Once a hazard has been identified, avoidance measures are sometimes taken too late – by which time a collision cannot be avoided. And the larger the ship, the longer it takes to reduce the vessel’s speed and take avoidance action, increasing the likelihood of a collision. Large cargo ships are difficult to maneuver, especially in the Bay’s narrow channels, and so they pose safety risks to recreational boaters in smaller vessels. A ship that is slowing down does not steer very well; the rudder needs a flow of water against its surface to remain responsive.
Collision avoidance technology
Collisions can be avoided by advanced technology and anti-collision systems. Collision avoidance for vessels highly depends on a robust obstacle detection. Marine radars are one of the most common navigation sensors used by marine vessels to detect other ships and avoid collisions. Radars can provide relative bearing and range of surrounding vessels over a wide area with a reasonable detection performance. Marine radars are X band or S band radars on ships used to detect other ships and land obstacles, to provide bearing and distance for collision avoidance and navigation at sea.
However, radar is not infallible. Systems have been known to function less effectively in more congested areas, and are susceptible to severe weather. However, the radar detection performance degrades when the sensor suffers from a blind zone at close range due to its sensing characteristics and environmental disturbances. Additionally, radars have a slow sampling rate and low resolution; thus their detection performance drops off when recognizing a high‐speed vessel or a small object protruding above the water surface. Lack of positive identification of the targets on the displays, and time delays and other limitation of radar for observing and calculating the action and response of vessels around, especially on busy waters, sometimes prevent possible action in time to avoid collision.
Therefore, additional sensors are necessary to enhance the autonomous situational awareness capabilities of Ships and USV systems in marine setting. In commercial ships, they are integrated into a full system of marine instruments including chartplotters, sonar, two-way marine radio, satellite navigation (GNSS) receivers such as the US Global Positioning System (GPS), and emergency locators (SART).
In fact, optical sensors such as lidars and cameras have been successfully used for many robotics applications, and they can be used to detect close-range obstacles in a radar’s blind zone at a relatively high sampling rate. Three‐dimensional (3D) lidars consist of multiple laser beams arranged vertically to collect surrounding obstacle information by rotating the lasers to achieve a full 360°‐environmental view. The sensors can be used to obtain precise relative bearing and range of close‐range obstacles in the radar’s blind zone. However, the effective sensing range is limited due to the relatively low angular resolution in the vertical direction, which brings few returns from a small‐sized obstacle.
On the other hand, cameras have relatively high angular resolution, and thus they can enhance the detectability of short‐ or mid‐range targets. The integration of these two sensors can lead to an ideal combination to improve the performance of target detection compared to that of the conventional radar‐only‐based approach.
Once a hazard has been identified, avoidance measures are sometimes taken too late – by which time a collision cannot be avoided. And the larger the ship, the longer it takes to reduce the vessel’s speed and take avoidance action, increasing the likelihood of a collision.
The information needed for safe navigation is currently obtained by combining radar data with visual information (unaided, binoculars, night vision). However, misjudgments accompanying visual observations comprise a major cause of ship collisions.
Furthermore, GPS signals are vulnerable to natural interference or intentional attacks, which can induce deterioration of signal reception by a GPS receiver integrated with AIS. Therefore, an additional sensing approach is required to detect surrounding ships whose motion information cannot be obtained from AIS equipment.
For automatic ship collision avoidance, the combined track data are used as obstacle information, and appropriate collision avoidance maneuvers are designed and executed in accordance with the international regulations for preventing collisions at sea (COLREGs).
Given the sensor measurements, the relative bearing and range information is extracted from automatic target detection algorithms and used to estimate the motion of targets in individual tracking filters. The tracking results are then combined in a central‐level fusion tracker. From the sensor‐level tracking filters, a data set of the estimated target motion (i.e., position, course, and speed) and their associated uncertainties are provided to a fusion‐based tracker. In addition to the sensor‐level track data, the motion information from AIS is used to update the combined track data in the central‐level fusion tracker. The autonomous collision avoidance system receives the estimated motion data from both the sensor‐level tracking filters and the fusion tracker and selects the tracking data to be used as obstacle information for automatic collision avoidance. An efficient collision‐free route is computed from the collision avoidance algorithm, and autopilot is applied to follow the computed route.
Automatic Identification System (AIS) Collision Avoidance Technology
In maritime traffic environments, the information of surrounding vessels can be obtained from navigational aid systems such as the automatic identification system (AIS), very‐high‐frequency radiotelephony, and the electronic chart display and information system. Among these, AIS significantly contributes to maritime traffic safety by providing maritime mobile service identity, position, course, and speed information of surrounding vessels with no need for much processing procedures.
The Automatic Identification System (AIS) is an automatic tracking system used on ships and by vessel traffic services (VTS) for identifying and locating vessels in real time. AIS technology identifies every vessel individually, along with its specific position and movements, enabling a virtual picture to be created in real time. The Automatic Identification System (AIS) is currently used on ships as a short-range tracking system and it is regulated by the International Maritime Organization (IMO). It provides the vessels and shore stations with information on identification and positioning on real-time in order to avoid ship collision accidents. AIS messages are transmitted using VHF radio waves and include: speed; position (latitude; longitude); course; heading; ship type; ship main dimensions; etc. AIS information is not degraded by rain-clutter like radar, so it works the same in all weather.
Class A transponders models operate at 12.5 watts, offering an average range of 30 nautical miles and are fitted aboard ships with gross tonnage of 300 or more tons. A vessel’s dynamic data, position, speed over ground and course over ground, is taken from the GPS. The Class A version uses SOTDMA transmission mode, giving it priority over Class B devices. Class B is applicable for the recreational and smaller commercial boat market.
Satellite-based AIS
Despite having been specified in the late part of the twentieth century it has only gained popularity over the last decade due to the use of satellite-based receivers which provides global coverage, improved response times and more reliability. Since 2008, satellites equipped with AIS receivers have been able to detect AIS signals transmitted by AIS transceivers on a global scale. Space-based AIS receptions open the possibility of unmanned transoceanic journeys, convenient for the transport of hazardous materials, which consequently enables the elongation of the duration of non-time-critical journeys, optimize the fuel consumption or even allows the direct use of electrical or solar power.
Additionally, these satellites serve as supplementary data sources for vessels and coastal authorities in busy port areas where conventional AIS receivers may not be able to cope with the large volume of ocean traffic. Satellite-based AIS provides an easy way for collecting AIS data on a global scale in almost real-time. Commercial exploitation of space AIS has been carried out during the last decade by companies such as SpaceQuest, Elane, ExactEarth, Marine
Traffic, ORBCOMM, and SPIRE
AIS data has become an important source of information for studying maritime traffic and associated risks. Computer programs have been developed for decoding and visualization of AIS data. The resulting image allows an analysis of the traffic, identifying the main routes and areas where the traffic is more or less intensive.
Most commissioned ships have an electronic chart information display system (ECIDS), radar, sonar and AIS graphic display. However, on some ships these three data sets are not integrated into a single display. The integration of these devices is very important as it becomes quite distracting to look at several different screens.
However, the AIS information is not always available because not all ships are equipped with the device (only regulated under certain ship conditions such as voyaging ships above 300 gross tonnage or passenger ships).
Autonomous COLREGs compliant ship navigation
The MAXCMAS (“MAchine eXecutable Collision regulations for Marine Autonomous Systems”) project aims at developing a COLREGs compliant path planner for autonomous vessel guidance and control. COLREGs are the “rules of the road” which were defined by the IMO (International Maritime Organization), to prevent collisions between two or more vessels. A significant challenge, which is tackled in the project, is to translate the COLREGs, which were written for human consumption, into state of the art collision avoidance algorithms.
MAXCMAS is a £1.27 million collaborative research project, with funding from InnovateUK. The project brings together key expertise of industrial partners: Rolls Royce (RR) as lead, Atlas Elektronik UK (AEUK) and Lloyd’s Register (LR); and academic partners: Queen’s University Belfast (QUB) and Southampton Solent University’s Warsash Maritime Academy (WMA).
The sensors information is fused, providing a world picture; the autonomy executive and collision avoidance algorithms generate navigation demands (heading, speed) to a controller interface. Those were then translated to control demands (throttle and rudder) for the autonomous vessel. The collision avoidance module (CAM) software regularly evaluates collision risks with the surrounding ships and/or landmass and, if necessary, provides collision avoidance decisions and actions that can be executed by the autonomous vessel.
Based on data provided to the CAM by the Autonomy Engine, if a target ship is present, the risk assessment submodule is activated which determines if there is a risk of collision with the target. To assess a risk of collision, the widely-used closest point of approach (CPA) method has been adopted.
AI can help in avoiding ship collision
Japan has an average of 286 ship collisions a year, varying in severity. Recently, Fujitsu has been testing AI technology in collaboration with the Japanese Coast Guard to calculate the risk of collisions and near-misses from traffic control rooms. According to the company, less experienced operators can quickly use the technology as effectively as their more seasoned peers.
“AI and big data technologies developed by Fujitsu Laboratories Ltd are used in the vessel collision risk calculation model. Using risk values calculated by Fujitsu’s technologies, operators can proactively detect vessels at risk and prioritise them. This will help in preventive planning while offering accurate information to vessels,” says Hiraku Fujimoto, manager of systems division IV, social systems unit at Fujitsu Limited.
“The challenge is to further enhance technologies such as the real-time processing of a collision risk calculation model based on automatic identification system (AIS) data. If we can do so, we can study the feasibility of combining the vessel traffic service (VTS) system with our technology and strengthen the collaboration.” “Fujitsu was able to learn about the difference in traffic conditions between Tokyo Bay and the Strait of Singapore, where we conducted a field trial in 2018,” adds Hiraku Fujimoto. “We also learned the necessity and method of tuning the collision risk calculation model suitable for each sea area.”
Orca AI’s AI-based navigation and vessel tracking system
While most cargo ships carry security cameras, computer vision cameras are rare. Tel Aviv’s Orca AI, a computer vision startup has developed an AI and computer vision-based solution that can be retrofitted to cargo ships and improve navigation and collision avoidance. Orca AI hopes its solution could introduce autonomous guidance to vessels already at sea.
Orca AI’s AI-based navigation and vessel tracking system supports ships in difficult to tricky to navigate situations and congested waterways, using vision sensors, thermal and low-light cameras, plus algorithms that look at the environment and alert crews to dangerous situations.
There are more than 4,000 annual marine incidents, largely due to human error. The company says this is getting worse as the coronavirus pandemic makes it harder for regular crew changes. The recent events in the Suez Canal have highlighted how crucial this industry is.
The company was founded by naval technology experts Yarden Gross and Dor Raviv. The latter is a former Israel navy computer vision expert. Customers include Kirby, Ray Car Carriers and NYK.
On the raise, Yarden Gross, CEO and co-founder said: “The maritime industry… is still far behind aviation with technological innovations. Ships deal with increasingly congested waterways, severe weather and low-visibility conditions creating difficult navigation experiences with often expensive cargo… Our solution provides unique insight and data to any ship in the world, helping to reduce these challenging situations and collisions in the future.”
Zohar Loshitzer, principal from OCV added: “Commercial shipping has historically been a highly regulated and traditional industry. However, we are now “witnessing a positive change in the adoption of tech solutions to increase safety and efficiency.