How the Mayflower Autonomous Ship Uses Edge Computing and Live Video to Conduct Vital Ocean Research

7 min read

The Mayflower Autonomous Ship has demonstrated the true power of edge computing by providing a platform on which oceanographic data can be collected across all the oceans to improve marine research.

As it prepares to make its transatlantic crossing, the Mayflower Autonomous Ship (MAS) will demonstrate the power of edge computing to help bring a whole range of capabilities to places that were previously out of reach — deepening the disciplines of oceanography and ecology, innovating to make us safer and more productive, helping better protect personal and private information and striving to get closer to fulfilling our obligations as custodians of a healthier and more vital world to cohabitate.

Built and operated by ProMare and Marine Ai, the mission of the Mayflower project is twofold:

  1. To experiment and refine the technologies needed to power a fully autonomous maritime vehicle navigating the far reaches of the world’s oceans.
  2. To create a platform on which huge volumes of oceanographic data can be collected to enable significant improvements in marine research.

Edge computing can provide greater access to precisely the type of innovation and power needed to unlock both of these key imperatives.

Central to achieving this are a set of technologies designed to harness that power — including things like video (and other sensor) capture and analytics based on artificial intelligence (AI), machine learning, decision making, intelligent automation and weather predictions – all of which are employed on the MAS research vessel.

But let’s take our own journey, delving deeper into some of the key edge technologies deployed on the Mayflower.

Low-bandwidth live video streaming

To safely navigate across an ocean and explore the sea, the Mayflower Autonomous Ship (MAS) relies heavily on video data as one of its key inputs. In addition to ocean exploration and detecting and identifying the objects it needs to negotiate in its path, the MAS must also support 24/7 live onshore video monitoring to see what the ship sees in the deep waters of the Atlantic Ocean. Yet, as any mobile device user might wonder, how is it possible to reliably stream video so far offshore?

Prior attempts by others to stream live video from platforms like the MAS had failed due to the combination of low, unreliable network bitrates combined with the enormous amount of data generated for analytics on standard CCTV cameras. Additionally, those systems were not suited to the transmission environments typically encountered out on the open seas.

In searching for a live video streaming partner, the Mayflower team turned to Videosoft’s solution, which is designed specifically for bandwidth-efficient video streaming applications of both Internet of Things (IoT) and closed circuit TV. With its efficient compression of video and reliable low-latency transmission over a variety of bandwidth constrained and unreliable networks (from as low as 6 kilobits per second), Videosoft (VS) provides an easy-to-use, integrated solution with adaptive video compression features and bespoke real-time transmission protocols.

On the MAS, the edge video processing manages all issues 24/7, including the local recording of high-quality footage and AI event-driven recording and database management. Only when required and appropriate for the ship to connect to central systems will it make those connections and provide a link to onshore-based services, including live viewing of the transcoded low bitrate video, downloading of high-quality recorded footage or moving cameras with pan and tilt (PTZ) control. This could not happen in a wireless architecture if the system was totally centralized, and it is an example of how moving processing and computing to the edge is necessary for the success and growth of remote and distributed applications.

During its voyage, the MAS can access multiple networks (such as satellite and cellular), depending on the proximity of the ship to these networks. The VS streams can run over either of them – the only difference being the latency of the network and the bandwidth available for video transmission. Using this technology, the MAS can reliably transmit its video streams, even across the challenging network conditions encountered across virtually all the ocean’s surface. Videosoft’s solutions can run in unattended operation mode to automatically respond to bandwidth changes and even self-heal after disconnections and reconnections or power down events.

The following diagram depicts the video system architecture – both onboard and on shore:

The following diagram depicts the video system architecture – both onboard and on shore:
  • Onboard: While recording is happening, the VS Gateway process is also communicated to shore-side servers to provide access to the stored data and video. This is always established whenever power is applied and comms are available to the ship. However, very little data is transferred by default.
  • On shore: When Marine Ai personnel on-shore need to view live video, they make a request to do so through client applications, at which point the video from the cameras is transcoded in real-time using Videosoft’s efficient compression techniques and then sent to the shore-side server on demand.
  • Online: To allow a public view of the MAS’ voyage, video streams are fed directly to Videosoft client applications and to the IBM Watson Media platform to be broadcast on the MAS400 website. Here's a live video clip showing the ship's dolphin escort.

Videosoft’s commitment to the MAS and its mission along with its track record of solving these problems has been a game-changer for the Mayflower project.

The video clip below was captured by Videosoft showing the ship's dolphin escort:

A new kind of edge computing platform for ocean science

In many ways, the Mayflower is the quintessential edge device – a “purpose-built machine with embedded compute,” to quote the formal definition. To perform its autonomous navigation tasks, the Mayflower must operate independently — relying only on the compute capacity that it has on board. To that end, the Mayflower has two Jetson Xavier NX systems and several Intel-based computers running Red Hat Enterprise Linux to host and support the software that enables Mayflower’s capabilities.

In fact, there is a lot of software on board the Mayflower covering everything from the video data processing we discussed previously, the object recognition software that feeds the navigation system, the propulsion system and its accompanying monitoring systems and the scientific payloads that we will discuss briefly later.

Early in the design of the Mayflower, the team had to deal with the question of how to manage the orchestration, deployment and optimization of the software components. For that, we turned to IBM Edge Application Manager (IEAM). IEAM is a commercial distribution of the Linux Foundation Edge (LFEdge) Open Horizon project. Its primary capability is to manage the process of deploying containerized software on to edge devices (and clusters) in the presence of diversity and dynamism — and to do so securely and at massive scale. It has turned out to be ideal for managing the deployment of software onto the Mayflower, managing the deployment of software to the wide range of devices around the Portsmouth Sound from which images were captured to train Mayflower’s object recognition software and emulating the Mayflower boat when testing the software systems.

The key benefit of using containers is that they are lighter and more granular than virtual machines (VMs). This has enabled different software components to evolve rapidly at their own pace and independent of all the other software components that make up the system. This dramatically simplifies the systems integration problem, removes bottlenecks in the build process and, thus, unlocks software development to iterate rapidly. It also increases the reliability of the overall system by isolating the individual components, reducing the number of variables each component must address. Container-based software development has become the ‘gold-standard’ of software engineering practices for cloud-native development, and all of those benefits commute directly to edge computing.

IEAM uses deployment policies to constrain what software to deploy and how. It avoids the need for administrators to make fine-grained decisions (a potential side effect of fine-grained software architectures) about what software to place where. Instead, IEAM will automatically determine the best placement of software based on the deployment policies set by the administrator at a high level. More importantly, IEAM continuously monitors the environment for any changes that might affect those deployment decisions and will automatically negotiate new placement decisions if the environment changes to a point of violating the original deployment policies.

This has enabled very rapid systems engineering processes whilst maintaining a very robust and stable software environment on the boat.

Enabling marine research at the edge

Although the oceans are the biggest feature of our Earth, they are, unfortunately, one of the least explored and understood. And, at the same time, our oceans play a huge role in the human impact equation – both serving to help absorb our impact, but also being detrimentally affected by our footprint. To do a better job of protecting our seas and, by extension, the Earth on which we live, the science equipment aboard the Mayflower is also a platform on which groundbreaking marine science can be conducted; more specifically, from which raw structured and unstructured data for marine science can be collected. In fact, the Mayflower is designed to hold up to 700KG of scientific equipment. In that space, we have deployed scientific payloads to sample CO2 levels, ocean acidification, algae, dyes and pollutants, phytoplankton density and marine mammal populations in the water along the journey. Open access to this invaluable data will be made available to marine scientists on shore to further study the health of our oceans, marine life and ecosystems, and how it is responding to global climate changes.

A big bet on technology for good

Frankly, this is what makes the Mayflower project so compelling. This is not just a commemoration of the past. Rather, it is a celebration of our future and sustainable developments that will be shaped by pioneering new approaches that will enable us to tread more carefully, reduce our impact and ensure marine conservation and the sustainability of our world. We collectively hope that the Mayflower project will be a foundational use case and noted in no small way as a turning point that will help ensure the viability of our planet Earth when historians from future generations look back at this moment 400 years from now.

Learn more

Learn more about the Mayflower Autonomous Ship:

Be the first to hear about news, product updates, and innovation from IBM Cloud