Setting Course, an ABS Podcast

The Future of Unmanned Vehicles at Sea with Saildrone

• American Bureau of Shipping • Season 1 • Episode 28

Share episode

In this episode of Setting Course, an ABS Podcast, we dive into the world of unmanned surface vehicles (USVs) and their impact on maritime operations — from commercial to research to national security. 

Featuring insights from retired U.S. Navy Vice Admiral and current Saildrone President John Mustin, ABS Senior Vice President of Global Engineering Gareth Burton, and ABS Director of Government Engineering Jihed Boulares, the discussion explores the key elements of successful autonomous systems and the critical role of human oversight. 

The guests also highlight the rigorous standards and rules already helping to enable safe and efficient operations for USVs in real-world conditions while regulations continue to develop.

Don’t forget to share this episode on social media, leave a review on your favorite podcast platform or send feedback to podcast@eagle.org. Discover how ABS is helping advance the maritime industry at www.eagle.org.

KEY POINTS

  • Saildrone’s USVs have achieved full classification status from ABS, setting a precedent for autonomous technology.
  • Effective risk mitigation for USVs relies on robust perception, reliability and trust to help ensure safe operations.
  • The regulatory framework for USVs is evolving, necessitating adaptations to existing guidelines for autonomous vessels.
  • Human oversight is essential in high-risk operations involving USVs.
  • The integration of autonomy in maritime operations is already happening, enhancing commercial operations, climate research, domain awareness and more.

GUESTS

A former U.S. Navy Vice Admiral and C-Suite technology entrepreneur, John Mustin brings to Saildrone more than three decades of experience at the intersection of national security, emerging technology, team development, and enterprise growth. A frequent speaker, published author, early-stage start-up investor, and advisory board member, he has led large, global enterprises and managed multi-billion-dollar budgets. With a BS in Systems Engineering from the United States Naval Academy, an MS in Operations Research from the Naval Postgraduate School, and an MBA from Babson College, he is an advocate for military and veteran causes and speaks and writes regularly on national security topics and post-military service hiring practices.

Gareth Burton is Senior Vice President, Global Engineering for ABS. He is a long-time ABS employee, having joined in 2001, as an engineer. While at ABS, Gareth has held several key positions, including Principal Engineer, Director of Engineering, Vice President of Operational and Environmental Performance, and Vice President of Technology, just to name a few. His positions have spanned the globe, including the U.S., Mexico and Singapore. In his current role, Gareth leads the ABS Engineering function, responsible for the review and approval of designs of marine and offshore assets for compliance with ABS, national and international requirements.

Jihed Boulares is the Director of Engineering of the Government Engineering Services group at ABS. He oversees design review of U.S. and international government vessels for compliance with ABS Rules and international regulations. Boulares worked in the maritime industry for 21 years in various roles, including naval ships operation and maintenance, naval ships structural design and analysis, and operational and environmental performance. He joined ABS in 2008 and has served in various engineering and leadership positions in the U.S. and Europe. Prior to ABS, he has held positions as surface warfare officer aboard a combatant ship, engineering duty officer at a naval shipyard and naval architect at McDermott International.

Brad Cox (00:07)
Welcome to Setting Course, an ABS Podcast, where we're charting the future of the marine and offshore industries. I'm your host Brad Cox and today we're going to be exploring the very interesting world of unmanned surface vehicles or USVs. 

We've certainly talked about autonomous systems on this podcast before, usually in the context of navigation assistance, collision detection and avoidance, and how the technologies will impact commercial vessel crews. Today's discussion will be a little different because we're talking about a fully autonomous technology that's already on the water supporting several applications. 

I'm talking about Saildrone, the maker of a range of fully autonomous deep water USVs. Saildrone’s 20-meter Surveyor vehicle recently received full classification status from ABS, while its smaller Voyager vehicle achieved ABS class in 2023. 

Joining the show today is retired US Navy Vice Admiral and current president of Saildrone, John Mustin. It’s a pleasure to have you on the show today, John.

John Mustin (00:57)
Well, thank you, Brad. I couldn't be more honored to join you.

Brad Cox (01:00)
And I'm also pleased to welcome ABS Senior Vice President of Global Engineering, Gareth Burton, to the show. Thanks for joining us, Gareth.

Gareth Burton (01:07)
Thanks Brad. It's great to be on the podcast to talk about the exciting topic of autonomy and where this is going within the marine industry.

Brad Cox (01:14)
And it's also great to have Jihed Boulares, Director of the Government Engineering Team at ABS, joining us. Thanks for being here, Jihed.

Jihed Boulares (01:20)
Thank you, Brad. It's pleasure to be here and to be part of this important conversation. The evolution of autonomous technology, especially in the marine industry, is something ABS is deeply engaged with, and I'm excited to share our perspective today.

Brad Cox (01:32)
So, let's jump right into the question I think a lot of people are going to have when it comes to uncrewed fully autonomous vessels on the water. And of course, this is a topic that's at the core of everything we do at ABS. 

So, John, starting with you, what are the risks associated with USVs? And how are they mitigated? I mentioned collision avoidance earlier. Is that something Saildrone’s vehicles can do?

John Mustin (01:52)
That's such a great question. I would just offer that when we talk about risk with uncrewed maritime systems, we really talk about two different dimensions. So, the first is the regulatory environment. And that defines many things about the way the systems operate within what's been legislated. But more interestingly, I would rather talk about the important part of the performance of the systems themselves. 

So given those two components, let me talk about system performance first. And I would say at the systems level, there are three pieces that we consider to be absolutely fundamental. The first is perception. The second is reliability. And the third is trust. 

So, perception in our world is all about what the vessels can see and how they see the world around them. Our robust perception models are used to detect, classify, and interpret what's on the water.

And so, you may ask, so how do you get more accurate perception models? I mean, it's a great question, but the answer is simple. You have to feed them an incredible amount of real-world data. So, we've sailed over 2 million miles in 50,000 days at sea. And during that period, we've been collecting the world's largest data set of marine imagery. And so, we feed that into our machine learning algorithms. We hand tag them so that we're able to add those to the classification and identification algorithms that make this one of the world's most robust and mature systems. So, we're very confident, frankly, about our ability to perceive the world around us. 

Then the second piece I mentioned before was reliability. I mean, our systems have to behave as expected and we need to be able to maneuver predictably. We need to maintain resilient communications. And frankly, we need to execute with precision those anti-collision avoidance algorithms that we described there. So we've got a lot of track record now in many, many days that have allowed us to show not only how we do it, but how well we do it. 

And then that feeds into that third piece, which was trust. Ultimately, the operator, and in our case, often that's a war fighter. Our customers tend to be the Department of Defense or the Department of Homeland Security, but they have to know that the platform is going to deliver as advertised. And these aren't controlled experiments. We're talking about real-world operations, many times in contested waters with national level stakes. We're interacting with other mariners. We use VHF, just like a crewed vessel would. And in many ways, that's what makes Saildrone unique. 

Our vessels have operated continuously with the U.S. Navy for years. We've been building an operational record that demonstrates those three elements specifically, it's perception, collision avoidance, reliability, and then trust at scale. And that's a long answer, but I would just tell you that, so we routinely identify those elements that we can iterate on in terms of how we perceive the world and how we react to the world. So anyway, we've got a great track record that we put up against any other autonomous vessel company to say, we do it better than anybody.

Brad Cox (04:53)
Thanks, John. So, Gareth, I'd like to get your thoughts on risk and safety with USVs. What are the key considerations as these sorts of technologies are being developed?

Gareth Burton (05:02)
Yeah, Brad, when we think about this technology, one of the key things we are thinking about is the specific application of it and the risks that are associated with that. You know, the risks associated with near shore commercial shipping operations differ significantly from that associated with deep sea or scientific type applications and therefore that drives the specific risk exposure.

For this reason, the ABS framework that we have established for autonomous technology is based on the concept of operations, or CONOPS. This presents the overall picture of the operations of the vessel and the operating environment in which it will function. 

So, some of the key risks associated with unmanned surface vessels would be associated with obviously the collision risk aspect. And as a result, it's really important to have many of those items that John was discussing. It's important to have reliable sensor systems and autonomy protocols in place. Likewise, compliance with items such as the COLREGs, the collision requirements are important for any other vessels that may be in the area so that they can identify those unmanned vessels.

Another risk would be associated with the autonomy and the autonomy protocols that are in place. Failures or shortcomings in that decision-making algorithms could have pretty significant consequences in the process. So making sure that those are fully established and validated are of key importance.

Brad Cox (06:32)
So, you mentioned the COLREGs, kind of branching off of that is the regulation discussion. So, there's COLREGs around preventing collisions at sea. And I know the IMO is working on a MASS code for autonomous ships. So, Jihed, what are the current regulations around USVs, if there are any, and what still needs to be done to develop those for this industry?

Jihed Boulares (06:53)
Yes, Brad, that's an excellent question. So the regulatory landscape for unmanned service vessels is indeed still evolving. Traditional frameworks like COLREGs, like you mentioned, are applicable, but as you know, were originally designed for crewed vessels. So there is a pressing need for more tailored guidance to address this complexity. 

You mentioned the IMO MASS code — this is a good step forward for sure. But it remains in its early stages. The non-mandatory version of the code is scheduled for 2026, while the mandatory code isn't expected until 2030. Meanwhile, some national and regional authorities are developing some interim policies and guidelines, but these tend to be very generic. 

At ABS, we are actively working to bridge this gap. So, since 2022, we developed the ABS Requirements for Autonomous and Remote Control Functions, which provides the industry with technical and survey requirements for vessels equipped with such capability. So this document basically established as a goal-based framework to guide the implementation of autonomous and remote control functions. 

And along with the other ABS Rules and requirements for standard hull systems, electrical, machinery and stability, it did form a clear framework for classing autonomous vessel. That's what we did with Saildrone. 

As for what still needs to be done, there is a lot to be done. This is an emerging technology that evolves quickly and we are learning as we go, too. So to meet these ongoing challenges, ABS is currently developing a comprehensive guide for unmanned vessels and aims to create a standard technical and survey requirement for USVs that will encompass not only vessel design, which is basically your traditional hull structure, stability system, machinery, but also a specialized USV element and autonomous and remote control functions. 

So this consistent focus will help ensure USVs are developed safely and efficiently while actually enabling innovation to thrive. That's very important to keep that in mind, too.

Brad Cox (08:48)
So it kind of sounds like the class requirements are the tip of the spear right now. John, I'd like to know what Saildrone’s experience was like with meeting those requirements.

John Mustin (8:59)
Well, it's no surprise that the current international regulatory framework for maritime vessels wasn't created to take into account the development of USVs, but Saildrone, not surprisingly, operates the USVs in good faith and we adhere as closely as practicable to all relevant provisions of both the United Nations Convention on the Law of the Sea and also the COLREGs, as Gareth mentioned too. 

And we do this with the understanding that the law was written for crewed vessels. And so we simply can't comply with every element of them, but where appropriate, we certainly comply with the intent. So we combine our automated collision avoidance systems and other systems with human in the loop oversight, and that ensures that we provide multiple layers of safety protocols. 

And candidly, we've had incredibly positive experiences engaging with both the U.S. Coast Guard and with you all at ABS, as well as international regulatory bodies. What we've always found to be true is the key is transparency. We want to prove our technology works reliably. We want to demonstrate that our operational protocols meet or exceed the traditional safety standards. And we also want to work collaboratively to ensure that the future definition is something that we can shape where appropriate. 

It's a learning process on both sides, but I've been encouraged by all the regulators' willingness to engage and adapt. And certainly we're all ears when it comes to ensuring that we're leading on that front.

Brad Cox (10:24)
So John, I'd like to go into a little more detail on how Saildrone does what it does. Could you provide some insight into what they do, how they operate, how long, and how far they go?

John Mustin (10:33)
Absolutely. So, Saildrones are the world's most advanced unmanned fully autonomous systems. And what we do is we provide governments, defense organizations, and commercial clients with persistent, precise maritime capabilities. And ultimately that really provides an advantage that allows our customers to either defend sovereignty, to secure commerce, or deter adversaries on the world's oceans. 

We currently have three sizes of vessels. Our smallest vessel is called an Explorer, it's a seven-meter vessel. We have a Voyager, that's a 10-meter vessel. And then our Surveyor is a 20-meter vessel. And as we increase in size, we include increasingly large and complex payloads, which address a range of different kind of maritime missions. But ultimately the two primary missions for us are intelligence, surveillance and reconnaissance and survey. 

And our vessels operate for months at a time without crews are refueling. And we measure the range in thousands of miles. I mean, frankly, because of our unique design that includes both wind assist and engine power allows us to have routine missions that exceed 3,000 miles. And our smallest vessel, the Explorer, routinely stays out for a year at a time. Our medium and large vessel, the Voyager and Surveyor, we measure in about 100-day missions. And that's simply a reflection of the desire to periodically top off a gas tank if necessary. 

Again, our unique design allows us to leverage the wind so that we don't use the engines unless we absolutely have to. And that's a big differentiator relative to other autonomous systems that maybe kind of, we call them go fast boats because that limits their range, it limits their endurance. And we are the beneficiaries of extreme endurance as a result of the thoughtful design that allows us to stay out of the water longer.

What kind of mission specifically? So, for bathymetric work, we can do bottom contour mapping down to 11,000 meters with our Surveyor, but we also provide this maritime domain awareness with 200-watt radar, electro-optical and infrared sensors. You know, we can stream full motion video. We've got satellite communications. 

There's certainly a variety of sensor packages that are tailored to the specific mission of our clients, but what we found is our platforms themselves are miraculous technological vehicles, but the real benefit of the service that Saildrone provides is the capture, transmission, and display of information. 

So, in a lot of ways, I say, while everyone loves our platforms, we're really a data company. And the way that we perform is through a couple of different business models. We can provide data as a service, as a mission as a service model, or we can sell platforms, provide training and allow a modicum of handoff during operations and maintenance that ultimately results in sovereignty with another customer that buys the boats and operates them themselves.

Brad Cox (13:37)
So, to take a step back and look at the bigger picture for the technology, a lot of our listeners are going to be on the commercial side of shipping. So, Gareth, are there any lessons the industry can take from these particular applications that could help the development of autonomous systems on larger commercial ships?

Gareth Burton (13:54)
Yes, Brad, there certainly are. And, you know, when we think about smaller vessels to larger vessels, the concept of operations are very much the same, or at least are conceptually similar. So the lessons that have been learned on smaller vessels can certainly be extended and scaled up to the larger ocean-going vessels. 

Some of those areas where we have seen lessons that can be learned that can be extended, would be on items associated with the verification and validation of the algorithms that are used in the process and the models behind those. And certainly the process that's used for that verification and validation process is equally applicable, no matter the size of the vessel themselves. 

Other challenges and lessons learned that are applicable would be items such as the data connectivity aspects, making sure that there's secure and reliable connectivity of that data transmission again would be equally reliable. 

The sensor fusion aspects of the various sources of data which are being collected and how those are brought together and used in the process. And then last but not least, the interaction between the conventional human aspects, where they appear in the loop, as well as the autonomous functions on board the vessel, and making sure that that interaction is as smooth and seamless as possible, again, is applicable across the size of the vessels.

Brad Cox (15:21)
So, you know, you mentioned the human in the loop, Gareth, how important is it to have some level of human oversight for these technologies?

Gareth Burton (15:28)
Yes, Brad, like many other items, it really depends on the risk of the operation being performed. You know, for some low-risk items, the deployment of autonomous technology really doesn't need much oversight. By contrast, some of the high risk operations, such as a fully autonomous vessel operating in a congested waterway, obviously the risk profile is completely different. And in that case, there would be a need for oversight and engagement with human in the loop.

The ABS framework that we have developed really looks at that and it’s really based on three different categories that we have established. The first category is the situation where if there is an issue, it would not lead to any dangerous situations for human safety, for the safety of the vessel or the associated environment.

Our category two would be the situations where a potential scenario could occur that could lead to those dangerous situations, again, for the humans, for the vessel, and for the environment. 

And then the third category would be the situation where a failure could immediately lead to a dangerous situation and potential impact to humans and threats to the environment. And based on these categories, then we set up the requirements associated for that human in the loop requirements.

Brad Cox (16:47)
John, can you provide a little insight into how you guys approach that oversight and human in the loop?

John Mustin (16:52)
Absolutely. So every one of our vessels, while capable of 100 percent autonomous operation, still has a pilot that is providing overwatch for every one of our vessels. So essentially, the pilot is responsible for mapping out a mission. And the mission then is performed via the autonomous operations of the vehicle. 

So, we, for instance, would create a plan of intended movement, a PIM track, and say, get from point A to point B, but it's over to the vessel to determine the optimal way to get there. And so they will harness the current, the wind, the environment to ensure that they do that the most effective way. 

As I mentioned, although we have both wind power and engine power, we only use the engine when we have to. Our preference is obviously to operate in accordance with the winds and currents when possible. And that just extends our range and our on-station time. So anyway, though, we always have visibility into where every one of our vessels is. That perception suite allows us to have great understanding of the other activities in the vicinity of our vessel. And that way we certainly can take control if necessary, but our autonomous algorithms ensure that that is never required.

Brad Cox (18:03)
I don't know how global this reference is going to be, it's probably an oversimplification of it. But it reminds me of the LEGO robotics, sort of STEM courses where you have to train it to do something. All the students watch and make sure it does what it's supposed to do. 

So, pivoting a little bit here. Of course, when we talk about sending something hundreds of miles off coast and operating on its own, I'm curious, how do you validate this kind of technology before it hits the water? Maybe this leans into the classification side of things, and we've been talking about this off and on, but Jihed, can you provide some insight there?

Jihed Boulares (18:37)
Yeah, Brad, validation is absolutely critical when it comes to autonomous systems, especially those designed to operate for extended periods without direct human oversight or little human oversight like the Saildrone in our case. 

At ABS, we take comprehensive and layered approach to help ensure these vessels can perform safely and reliably in real-world conditions. So, our process begins with rigorous design review followed by simulation-based testing and real-world trials actually, including the sea trial and even after delivering the vessel as it stays in class. 

But it goes far beyond traditional assessments. So, we evaluate structural integrity, machinery, electrical systems and stability, the core elements of classification that we do for any commercial vessels, while also diving deep into the unique aspects of autonomy. So, this includes reviewing the concept of operations, like Gareth said, conducting risk assessment and establishing test and survey programs. They are all tailored to autonomous functions. 

We also examine management of change protocols. Cyber resilience is huge when the vessels are operating in the middle of nowhere. And software testing plan to help ensure the digital backbone of the vessel is secure and dependable. Another critical area is data management. And I think John touched on data. We assess how data is collected, validated, modeled and recorded. This is essential for both operational performance and post-deployment analysis once the vessel is delivered. So all these elements come together to form a robust validation framework from ABS.

Brad Cox (20:08)
And speaking of that classification, John, why did Saildrone decide to pursue class for its vehicles? And what did it mean to you guys to achieve that?

John Mustin (20:17)
Well, I wouldn't be here if we didn't consider ABS class as the gold standard. So that's a true statement. And I'll just differentiate the classification standard from certificates or MOUs, which we consider to be generally statements of intent or preliminary validations. And I see that a lot with many other unmanned systems now where they're beginning to dip their toe into what it means to achieve class standardization.

But for us, we recognize that means ABS has formally recognized the vessel is meeting the same rigorous design, safety, and operational standards as a crewed ship. And furthermore, they've entered it into the official register. 

I like to remind people that we have almost 60,000 days at sea, which is the equivalent of a single ship sailing for 164 years. And so you can certainly understand that we can collect and learn and iterate quite a bit based on that experience. And so with all of that experience, we have a level of maturity and reliability that is obviously necessary to satisfy ABS' full review. 

So the great news for us is now we have two vehicles. So the Voyager and the Surveyor, each of which is fully classed by ABS. And that's the difference between signaling ambition and being recognized as a true seagoing vessel. So that underscores the extreme liability, the proven maturity of our designs, and in many ways, this classification sets the bar that we know all USVs will ultimately need to reach if they're going to be trusted for real world operations.

Brad Cox (21:53)
And Jihed, I feel like we've touched on this already between talking about regulations and validation, or I should say requirements, not regulations. But from the ABS perspective, how significant was this classification and what were some of the unique considerations?

Jihed Boulares (22:08)
So achieving classification for a USV like the Saildrone Surveyor is a major milestone for ABS. It's not just a stamp of approval. It's demonstration that the vessel meets a comprehensive set of safety and performance standards. 

What made this particular project unique was adapting our traditional classification framework to accommodate autonomous operations, alternative propulsion, and remote monitoring too. It requires collaboration with Saildrone. It requires innovation and, to be frankly, it requires a forward-looking mindset. 

For ABS perspective, it sets precedent for how we support the safe integration of autonomous technology into the broader marine ecosystem. It also helps us to test our autonomy Rules and at the same time, it helps us consistently improve these Rules as we go.

Brad Cox (22:56)
So, guys, this has been really great. As we kind of get to the end of our time here, I wanted to go ahead and give each of you a chance to provide some closing thoughts, maybe a key takeaway for our listeners or maybe even a bold prediction for the future. John, you want to start us off with that one?

John Mustin (23:10)
Absolutely. I would just start by saying unmanned is not the future. Unmanned is today. It's here and wise entities are taking full advantage of what autonomy provides. 

What we're seeing now is a hybrid fleet development in the defense industry. We're seeing commercial systems that are supplemented by unmanned and autonomous vehicles. And there are just so many benefits, some of which we've talked about — notwithstanding the risk mitigation, the cost efficiencies, the speed for completion given that ships don't need to come back for maintenance or fueling. There's so many benefits that make this a very compelling business case that I believe we're at a tipping point where the adoption and understanding of autonomy's impact on the world's oceans is exploding. 

And that's obviously very exciting to us as we begin to incorporate increasingly complex payloads, sensor suites, and mission requirements. So, you know, we are very excited every day to engage with our clients and say, what else can we do? What's next? You know, we have always viewed the world through the windshield more so than the rear view mirror. And so I would tell you that, you know, our great interest is in helping to shape and define what the future of maritime autonomy looks like.

Brad Cox (24:32)
Okay, thanks. And Gareth, any closing thoughts?

Gareth Burton (24:35)
Certainly wise words there from John on the topic. You know, as the industry continues its journey towards embracing autonomous systems, it's important to remember that these technologies should be used as tools, tools to enhance safety, efficiency, and sustainability. And again, John had touched on many of those during his comments there. The success really lies in striking the right balance between increasing autonomous functions and that human oversight that we talked about before. And really the key is making sure that there's no unintended safety consequences in the overall process.

Brad Cox (25:14)
Okay, great. And, Jihed?

Jihed Boulares (25:15)
So I completely agree with John. I think the key takeaway is that autonomy in maritime isn’t in the distant future. It's happening now and we can see it from the Saildrone vessels that are currently classed with ABS. 

But to scale safely and effectively, I agree with Gareth, though, we need robust framework for validation, regulation, and classification. 

As for a bold prediction, I believe we'll see autonomous vessels playing a central role in climate research, ocean survey, military is going to be a big application and even commercial just within the next decade.

Brad Cox (25:48)
Okay, great. So, to wrap us up, it was great to dive into USV technology, what it means for the maritime industry, mitigating risk, and ultimately why class matters for these types of vehicles. So John, Gareth, Jihed, it was wonderful to have you join the show for this discussion, and I look forward to an opportunity to get together again in the future as the technology advances.

John Mustin (26:08)
Well, many thanks for the gracious invitation. It was a pleasure to be on the panel with the two gentlemen here.

Gareth Burton (26:13)
Thank you, Brad.

Jihed Boulares (26:14)
Thank you, Brad.

Brad Cox (26:15)
And for everybody listening, thank you for joining us for another episode of Setting Course. Be sure to subscribe, leave a review, and share this episode. To keep up with the latest technologies reshaping the maritime industry, visit us at www.eagle.org. Thank you for listening.