The maritime industry can use an edge AI appliance such as Lanner’s EAI-I130 deployed on the boat to bring automation and intelligence to the boat (network edge). This edge AI appliance collects the data generated by sensors attached to the vessel and then runs AI inference.

Challenges/requirements

Technologies like automation and AI/ML have been mainly used in inland industries. On land, the computing, storage, network, and power infrastructure necessary to back these technologies is more accessible. After all, these technologies heavily rely on centralized cloud computing, so Internet access is paramount.

But on water or deep in the ocean, the access to these resources is an entirely different story. The following are the challenges when introducing automation, AI/ML, or analytics into ships or vessels:

a. Internet connectivity. Ocean-based industries, including maritime and shipping, have limited access to Internet connectivity. Boats and ships usually operate in remote offshore waters where access to the Internet is low to zero. The wireless network connectivity in these places is unstable, unreliable, and most likely inexistent. Such inconsistent and unavailable communications not only put the crew and passengers at risk but also makes access to cloud computing impossible.

b. bandwidth. Modern vessels or passenger ships like cruises may switch to land-based mobile communications while near land and switch to VSAT communications while traversing the open seas. Although boats and ships could still use modern VSAT communications (with large bandwidth), the large amount of data needed for any AI inference, ML, or deep learning algorithms can quickly overwhelm any satellite bandwidth.

c. Power efficiency. All boats and ships have limited access to power resources. They may use batteries and solar energy to power their electric appliances. Ships and vessels are not connected to any electrical grid, so the power consumption on additional electronic appliances must be lean. Electric and electronic devices deployed on boats to run 24/7 must be designed with minimal power consumption.

Edge AI Solution

Edge AI is the term referred to when AI is deployed on a "local" edge network, as opposed to a centralized cloud-based computer. It allows an always-on and reliable AI solution— where all AI inference happens at the edge. Edge AI removes the risks associated with losing internet connectivity, high latencies, or delays while connecting to the cloud for AI. Although a boat or ship could still use an Internet connection to upload collected data to the cloud for further AI processing, or monitoring, with edge computing (robust computing and storage), AI could also work offline. With edge computing, connectivity and bandwidth are no longer issues. As the local edge appliance can process data effectively, run AI, and make decisions.

Ferry boat vector created by macrovector

Components of edge AI for the maritime autonomy

  • Data collection: IoT devices and sensors can collect a breadth of valuable real-time data from the ship. A few examples of collected marine telematics data range from weather and visibility, wind, AIS data, GPS, radar, Solid state compass, engine monitors, fuel monitors, etc. Cameras may also play an important role in collecting imaging data for Computer Vision applications.
  • Edge AI appliance. An edge AI appliance (such as EAI-I130) collects the data generated by these IoT sensors or cameras deployed around the boat or ship and runs AI inference. This appliance is installed on the boat and allows on-device data processing (AI, CV, or data analytics) for off-grid applications.
  • Network. A fast built-in Ship Area Network (SAN) can be used to transmit all sensors and cameras to the edge computing appliance. An IoT gateway can help connect any IoT and other devices to the Internet. Although Internet connection in ships and boats is limited, it can still empower edge computing for further analysis and monitoring.


Talk to Us
 

The Edge AI Appliance: EAI-I130

Lanner’s EAI-I130 (Edge AI) is an industrial-grade AI inference system designed for the edge. The appliance can handle multiple AI workloads and consumes very low power. The appliance is also compliant with IP40, comes in a fanless design, and supports robust wireless, including 5G and WiFi6. All qualities make it an ideal solution for maritime automation.

The appliance is powered by NVIDIA® Jetson Xavier NX— a series module referred to as the world's smallest AI supercomputers. The Jetson Xavier NX brings supercomputer performance to the edge in a small form factor (SOM) System-On-Module delivering up to 21 TOPS (Trillion OPerations/Second) of throughput. Jetson Xavier NX allows the EAI-I130 to maintain a small size and low power consumption while still dramatically increasing its AI capabilities.

With this power on the boat, the EAI-I130 can run AI workloads or computing vision and allow autonomous maritime applications.

EAI-I130 Highlights.

  • Robust wireless 5G, LTE, and WiFi6 support
  • IP40 standard fanless design.
  • Operates at -40°C To 70°C temperature range.
  • 384-core NVIDIA Volta GPU With 48 Tensor Cores.
  • 2x GigE PoE LANs With Support For IEEE 802.3 af/at PoE(+).

Benefits

The Nvidia Jetson-powered platform (EAI-I130) introduces edge computing to the challenging maritime industry. As stated above, edge computing solves the challenges of having to connect the boat to the Internet for AI and having to upload large amounts of data. The edge AI appliance takes care of everything AI while off-grid.

Below are a few applications of the edge AI Nvidia Jetson-powered platform for maritime autonomy.

  1. Smart shipping. Smart ships are autonomous ships that collect data from sensors and IoT devices so that they can autonomously maneuver or alert the crew to take further action. Since smart ships are highly dependent on real-time action, they can leverage edge computing or edge AI for real-time data processing. Smart ships have the extra pair of eyes to help monitor when the crew is distracted, watch for blind spots, or alert (and avoid) collisions.
  1. Autonomous Docking Systems. The maritime industry can also use the edge AI technology for marine autonomous docking and position systems. These systems can help a ship control its speed and provide a safe and precise path to deliver it to its dock (using predetermined navigation and docking procedures). An autonomous docking system is integrated into the ship and uses external sensors to provide information about surrounding objects, winds, and currents.
  1. Crew welfare and security. Although most of the safety and welfare of the crew would depend on having a reliable Internet connection, having additional computational power on the boat can also help improve security. For instance, AI and ML technologies help detect objects, analyze historical data, and draw conclusions and predictions. A vessel’s machinery or other telematics can be monitored by intelligent systems and take corrective actions or determine predictive maintenance before catastrophic accidents develop.

Featured Product


EAI-I130

Industrial Grade AI Inference System For 5G Edge With NVIDIA® Jetson NX

CPU 6-core NVIDIA Carmel ARM®v8.2 64-bit CPU 2MB L2 + 4MB L3
Chipset N/A

Read more