OPINON: Tackling the Houthi Drone Threat with AI
Houthi rebel drone attacks on vessels traversing the Red Sea and Suez Canal route, accounting for 10% of global trade, have exposed the shortcomings of current ship alert systems.
The unpredictability and increased frequency of these attacks have created a challenging security environment for ships passing the area via the Bab el Mandeb strait, posing a substantial risk to seafarers and maritime trade by endangering the safety of the crew and cargo.
Initially focused on Israel-related ships in solidarity with Hamas, the Houthi attacks have now escalated into a broader threat targeting any vessel in the region, with a US-led multinational naval task force, named Operation Prosperity Guardian, mobilised to counter the threat. So far, as many as 27 vessels - mainly container ships and gas carriers - making the precarious passage have come under fire with drones and ballistic missiles from Houthi rebels in Yemen since the attacks started on 19 November last year, according to the latest figure from US Central Command at the time of writing.
Impact on crew welfare
These new threats have an impact on several fronts: delays to the ship's schedule and damage to the vessel itself are only part of the problem. The impact on the crew navigating the vessel, who find themselves in the middle of an attack, is of much greater concern. Drones can strike a vessel without warning at any given moment from any direction, which can trigger fear, stress and panic among the crew and allow them little or no time to safeguard themselves. Conversely, being able to monitor incoming drones would allow early warning to provide precious time for crew to find shelter, attempt an evasive maneuver, or report an imminent attack.
Fortunately, there have so far been no seafarer fatalities as a consequence of these attacks, but the crisis is having a significant impact on maritime trade with major shipping companies including Maersk, MSC, and Hapag-Lloyd opting to avoid the region by rerouting their vessels on a much longer route around the Cape of Good Hope - negatively affecting fuel costs and emissions.
The Gulf of Aden saw a 40% decrease in ship arrivals in the five-day period through December 26 last year, with container ship arrivals plummeting by 87%, and LPG/LNG vessels declining by 30% and 28%, respectively, according to Clarksons data. Suez Canal transits mirrored this trend, with southbound transits down 45% and northbound transits down 26% during the period.
Flying under the radar
Traditional detection systems such as ship radar are limited in their ability to protect personnel, cargo and assets effectively. In incidents faced by shipping companies over recent months, existing ship technology was unable to track the trajectory of these small and lightweight drones to give crews sufficient time to react and seek safety.
Unlike missiles, the compact size of Houthi drones, with a wingspan of approximately 4.5 meters and a length of no more than 2.5 meters - similar to a sailboard - poses a significant obstacle for marine radar systems. Considering the height factors, marine radar is not designed to detect such small objects, rendering it ineffective in identifying these drones.
Moreover, the speed of these drones, moving at 200-250 km/h, exceeds the capabilities of marine radar, which typically operates with an average rotation cycle of 2.5 seconds. This limitation makes it challenging to track high-speed targets efficiently. The lightweight construction of the drones using materials like carbon fiber and aluminium further compounds the difficulty, as their low altitudes escape the detection capabilities of marine radar systems, and make them difficult to jam using standard anti-missile technology.
Combined with potential radar clutter, this makes it practically impossible to properly track this kind of target and understand its motion profile based on generated parameters such as distance, CPA and TCPA. Additionally, the inherent inability of radar to classify any target introduces the risk of confusion, making it challenging to differentiate between actual drones and unrelated elements such as sea clutter or clouds.
AI to counter the threat
AI-based target detection can play a crucial role in mitigating drone attacks on ships by enhancing their ability to detect and respond to potential threats. Here's how AI-based target detection can help address this security challenge:
• Early detection:
o AI-powered systems can continuously monitor the ship's surroundings, including the airspace and the surrounding waters, using a combination of sensors such as cameras, radar, and lidar.
o These systems can identify and track incoming drones or potential threats, even in low-light or adverse weather conditions, with a high degree of accuracy and speed.
• Anomaly detection:
o AI algorithms can establish a baseline of normal activity around the ship and identify any anomalies or deviations from this baseline.
o If a drone approaches the ship in an unusual or unexpected manner, the AI system can alert the ship's crew or security personnel to investigate further.
• Classification and identification:
o AI can classify detected objects as potential threats or non-threats by analyzing their size, speed, flight patterns and other characteristics.
o By using machine learning models, AI can also identify specific types of drones or unmanned aerial vehicles (UAVs) that pose a threat, helping security personnel assess the level of risk.
• Autonomous response:
o Once a potential threat is detected and confirmed, AI systems can trigger automated responses to neutralize the threat or deter it from approaching the ship.
o Possible responses include activating counter-drone measures, such as jamming communications or deploying physical countermeasures like nets, lasers, or even interceptor drones.
• Integration with existing security systems:
o AI-based target detection can be integrated with existing ship security systems, such as surveillance cameras, access control and alarm systems, to provide a comprehensive security network.
o This integration enables real-time coordination and communication among various security measures and personnel.
• Continuous learning and adaptation:
o AI systems can continuously learn from new data and adapt to evolving threats and tactics used by malicious actors.
o This adaptability ensures that the ship's defense mechanisms remain effective against emerging drone threats.
The AI-powered SeaPod platform, developed by Orca AI, is currently the only maritime tool available that can address this threat. The reliable technology excels in the early detection of small targets, notably even airborne targets, providing timely alerts to crews for actions such as taking cover or recording for evidence.
A notable feature is the platform's capability to deliver at least a one-minute specific audio warning to the crew before a potential attack, allowing them to undertake the necessary precautions to ensure safety. The platform also provides live video streaming for onshore monitoring and automatically adjusts itself to a specialized mode in geo-fenced regions like the Red Sea, enhancing its sensitivity dramatically.
The new reality requires us to consider a new security approach and broader adoption of new tools and security measures. It's important to note that while AI-based target detection can significantly enhance a ship's security against drone attacks, it should be part of a broader security strategy that includes legal and regulatory compliance, physical security measures, and well-trained security personnel. Additionally, the use of countermeasures to neutralize threats should comply with local laws and international regulations to avoid legal repercussions.