The attacks by rockets, skiffs, and drones on commercial cargo ships transiting the Red Sea by Houthi rebels in Yemen highlight the security dangers merchant vessels face on the high seas. In response, many, but not all, carriers announced that their vessels would no longer be transiting the area of hostilities, diverting them instead around Africa’s Cape of Good Hope.
For those still navigating the danger zone, deploying an early warning system onboard could help keep crews safe. A system powered by artificial intelligence, developed by Orca AI, a Tel Aviv-based company, could fit that bill, although it was developed with other navigational challenges in mind.
Raviv has spoken with fleet managers whose ships have come under attack in the Red Sea, and based on those conversations, he concluded that the missile threat is not their biggest concern. “The bigger threats are the skiffs and the drones,” he said, “which are invisible to existing systems of navigation sensors.”
Orca AI’s system can’t prevent the Houthi attacks, nor repel them, nor provide enough time for vessels to change course in response to an imminent attack. “But they can,” says Raviv, “detect these threats in time for the crew to get to a safe spot and safeguard their lives.”
AI Designed To Be Adaptable to Hazards
Orca AI may not have been developed with military dangers in mind, but it is adaptable to those situations because it was designed to use advanced sensors and AI to detect, and provide solutions to, hazards, such as whales, boats, and buoys, that vessels could encounter in their navigational paths. The system is now deployed on over 260 commercial ships worldwide. One of the company’s largest customers, Mediterranean Shipping Company (MSC), has the system deployed on, or on order for, 600 of its vessels.
Orca AI is designed to provide early warning of hazards that vessels may encounter, so that the crew can react sooner. Besides enhancing safety, this feature also introduces operational efficiencies by allowing crews to take actions that minimize fuel consumption.
The latest iteration of the system, known as Seapod which was released in 2023, includes five day-view high-resolution cameras and three thermal cameras with wide fields of view. “The new technology introduces distance calculated solely by computer vision and artificial intelligence,” said Raviv.
Computer vision describes the ability of algorithms using AI to use data captured by video sensors and other devices to infer a ship’s surroundings; to detect, track, and classify hazards; and to estimate their distance from the vessel. AI also provides a layer of context which considers the location of the vessel—such as whether it’s close to port or to shore or to no-go areas, and whether there are dangerous depths in the vicinity. Alerts on risks and recommendations for prioritization of action are based on the fusion of all this data, as well as on Orca AI’s own collection of data on bodies of water representing 60 million nautical miles from around the globe.
“One of the biggest challenges for ship navigation is when it's entering congested waters,” said Raviv. “Many of the rules of the sea are not enforced properly, so it's very hard to understand what is risky and what is not. Computer vision can understand context and to recommend alerts to highlight dangers.”
Orca AI’s technology could also contribute to the autonomous navigation of vessels, although Raviv does not expect that the shipping industry will accept fully autonomous vessels anytime soon. “There’s a difference between autonomous navigation and fully unmanned ships,” he explained. “Autonomous navigation is like autopilot in aviation, where you still have pilots on board. I do not think that we'll have fully unmanned commercial ships in the next years, because there is no real economic benefit other than crew reduction. But autonomous navigation technology will boom in the next few years, specifically around safety of navigation and recommendations and for emissions reduction and fuel economy.”
New Decarbonization Module
The company plans on introducing a new decarbonization and emissions reduction module this year which has the potential to save ship operators substantial costs. “Shipping companies don’t take into account the correlation between the timing of reactions to encounters at sea and the cost of fuel,” said Raviv. “We have learned that most crews react quite late, between 15 and 20 minutes before an encounter. This incurs a large deviation from the original route which increases fuel costs.”
Notifications to crews to alter course well in advance of encounters with hazards or other vessels allows for small course corrections, which save money. Says Raviv, “I'm talking about anywhere between $150,000 to $250,000 per ship per year on unnecessary fuel expenditure due to late action.”
