The Orca AI platform is made of three main units:
1. A Look-Out-Unit called “SeaPod”, made of 5 HD (day) and 3 thermal (night) cameras, which is usually installed on the compass deck, Fore-mast or the Christmas tree, covering 225 degrees day field-of-view and 100 degrees thermal field-of-view.
2. A monitor on the bridge for the crew and a monitor at the Master cabin – both constantly showing what the Look-Out-Unit sees.
3. A dashboard called “FleetView” for the operations team in the office.
The Orca AI platform works in three main stages:
1. It connects to the mandatory onboard navigation sensors.
2. It integrates them with high-resolution cameras and proprietary AI algorithms to alert the crew on dangerous targets.
3. It prioritizes risky targets in real-time and helps to make better decisions in complex navigation situations like congested waters and low visibility.
No, it doesn’t. Radar is the official collision avoidance tool regulated by SOLAS. The Orca AI platform is complementary to the radar and ECDIS and together with them provides full coverage for optimal situational awareness.
Yes, it is. In February 2022, the Orca AI platform received Product Design Assessment (PDA) certification from the American Bureau of Shipping (ABS) and became the world’s first AI-based navigation platform to receive this certification. In addition, Orca AI collaborates with leading-class organizations such as ABS, DNV and LR, to make sure our solutions are constantly reviewed and comply with regulatory requirements.
The Master of the ship is responsible for all actions and decisions, as with any other system onboard. Orca AI shall conduct constant visual detection and confirmation of marine targets, highlight risks, and prioritize them according to COLREG, yet discretion should be performed by the OOW.
The platform is connected to nine sensors on the vessel: AIS, Radar S-band, Radar X-band, GPS, Gyrocompass, Echo-sounder (depth), rudder angle, Anemometer (wind), and engine RPM. It fuses the information received from the ship’s navigation sensors alongside computer vision sensors and AI algorithms.
The SeaPod is installed on the highest point available such as the compass deck, Christmas Tree, or Fore Mast.
The SeaPod includes 5 high-resolution day cameras and 3 thermal cameras. Day cameras provide a FOV of 225° and thermal cameras provide a FOV of 100°.
Yes, it can. The platform can detect floating containers, fishing boats, docking and sailing vessels, motorboats, ferries, navigation hazards, buoys, and whales. The detection range depends on a number of parameters, including the Look-Out-Unit installation height, visibility conditions, size of the target, and ambient temperature.
In most cases, large marine targets can be detected during the day in high visibility conditions up to 8NM. In low visibility conditions, this range is reduced to 0.5-4 NM. In general, the range highly depends on environmental conditions, such as ambient temperature and humidity.
Yes, it is. Orca AI SeaPod houses low-voltage cameras (24V supplied by the vessel, which is converted into 12V). The system is installed in a safe zone on the compass deck, next to other electronic devices (e.g., Gyrocompass, radar, etc.), without compromising them.
This is a valid concern, but as safety mechanisms evolve, data shows the opposite is true. For example, when public awareness about seat belts grew, drivers did not drive more aggressively – the opposite, they drove with more caution.
The crew is using the system as a decision support system which is complementary to the existing navigational tools and they keep watching outside the bridge and acquiring targets in the Radar and the ECDIS.
The screens for the system onboard the vessel have minimal interference with the crew’s biological night vision. The platform allows the user to dim the screen, to dim the video and also provides a dark mode of the entire UI, which works better in low brightness.
Video Recordings are performed automatically according to thresholds set by the office in various congestion-level conditions. These videos are uploaded to the cloud and can be used for navigational audits, or additional investigation. All the relevant event information is accessible including event summaries, videos, and maps.
The Orca AI FleetView highlights the fleet’s safety and operations KPIs including: safety score trend, top events, distribution of events and recommendations. The user can explore individual events with a playback on ENC (Electronic Navigation Chart), video recording and additional even information. The following event types are constantly logged in the dashboard:
• Sharp maneuvers
• Heavy weather encounters via pitch and roll
• Close encounters
• Speed drops
• UKC violations
All events can be filtered according to congestion level, location, COLREG and more, in order to create the best-customized dashboard experience.
The Orca AI SeaPod onboard alerts the crew in real time, based on the settings set by the Master. Alerts to the office are based on threshold settings in the Orca AI FleetView, according to the fleet safety policy under different navigation conditions. Events such as extreme rolling and pitching, SOG compliance violations and no-go zones are sent automatically and immediately to the office.
Orca AI has developed a proprietary data pipeline, designed specifically for deprived bandwidth environments onboard. The data is constantly monitored by the platform visibility team, alongside tech-ops specialists and QA.
Data requirements are flexible according to the customer’s data plan. Most common bandwidth is between 300 Kbps upload and 1 Mbps download, where Orca AI’s data consumption impact on download is minimal, and limited to remote software upgrades only, pending the user’s approval.
Orca AI is certified by ISO 27001. The customer’s data is confidential and is kept in Orca AI’s secured private cloud (VPC). Data transfer is conducted via VPN (Virtual Private Network) and is encrypted both in transit and at rest. Orca AI only uses the data to train the algorithm in an anonymous and aggregated manner.
No there isn’t. The system creates information based on relative information received by the cameras and is not dependent on potentially spoofed sensors.
In early 2022, Orca AI performed the world’s first autonomous ship navigation for a commercial shipping vessel in Japan. And this wasn’t just setting the course straight on the open seas. Our autonomous navigation system helped guide a commercial cargo ship on an almost 500-mile voyage in Tokyo Bay, which is one of the busiest ports in the world. The voyage lasted over 40 hours. During the autonomous maiden voyage, Orca AI’s platform performed 400 maneuvers and avoided over 100 potential incidents with other vessels. The commercial cargo ship’s journey was super successful and we believe this is only the first step in a global move towards autonomous shipping vessels.
There’s no doubt about it, the oceans are more congested than ever with an increase in global demand for goods. Yet, at the same time, there’s a huge shortage of professionals in the shipping space. Shipping companies are now facing a situation where it’s becoming harder and harder to find skilled seafarers. In a reality in which less experienced crews are operating ships, automation can help reduce human errors and avoid safety events.
It’s already happening. But it’s important to note that autonomous commercial ships are not the same as unmanned shipping vessels. Orca AI is working towards this goal, but for this to happen en masse, there must be a natural progression from partial automation to high automation, and then eventually to full automation. Only after successfully navigating this course will the next chapter begin; namely, unmanned shipping vessels. This will take a few years, but we’re already seeing a huge interest in our automated shipping system. And this is fantastic news, because it will alleviate congested waters and make them a safer place.