Dear Deeply Readers,

Welcome to the archives of Oceans Deeply. While we paused regular publication of the site on September 1, 2018, we are happy to serve as an ongoing public resource on ocean health and economy. We hope you’ll enjoy the reporting and analysis that was produced by our dedicated community of editors and contributors.

We continue to produce events and special projects while we explore where the on-site journalism goes next. If you’d like to reach us with feedback or ideas for collaboration you can do so at [email protected].

Solving the Sky-High Costs of Ocean Exploration with A.I.

Research ships are vital for advancing marine science but are costly to operate. Oscar Pizarro, a scientist at the University of Sydney’s Australian Centre for Field Robotics and the Schmidt Ocean Institute, thinks automated expeditions are the future of ocean science.

Written by Ian Evans Published on Read time Approx. 6 minutes
The team on board the R/V Falkor begins early-morning testing by lowering a drone into the ocean.SOI/ Leighton Rolley

A major challenge in studying the ocean is simply the cost. According to a report published by UNESCO last year, ocean research vessels rack up charges ranging from $10,000 to more than $40,000 a day to operate. Increasing costs “will almost certainly” constrain the future use of such ships, researchers found.

One of the biggest expenses? The crew itself. That’s why Oscar Pizarro, a principal research fellow at the University of Sydney’s Australian Centre for Field Robotics and chief scientist at the Schmidt Ocean Institute, thinks taking humans out of the mix is an answer. He recently led a research cruise testing automated research drones that he hopes could, eventually, undertake an expedition on their own from an autonomous ship.

The Slocum glider from the University of Michigan. (SOI/ Monika Naranjo Gonzalez)

The 20-day expedition on board the research vessel Falkor returned this week after deploying a fleet of drones to map the seafloor off the coast of Hawaii. Teams from several research institutions and universities used a variety of drones, aiming to improve each automated system and the team’s ability to keep track of multiple drones at once. It’s an idea that Pizarro thinks will one day lead to cheaper, automated expeditions, leaving human researchers on land.

Oceans Deeply spoke with Pizarro about the expedition while he was aboard the Falkor, traveling from the coast of Maui to the Big Island.

Oceans Deeply: Why is it important to be mapping the coral reefs and the seafloor with automated technology?

Oscar Pizarro: We really depend, as a global civilization, on the ocean. We really need it to be healthy for us to also stay alive and healthy; and in order to make informed decisions and trade-offs between how we exploit it and how we look after the ocean, we need to have good science, and science ultimately relies on having good observations.

If observations are expensive, you will have fewer of them, and you will have to rely more on your assumptions and models. But if you have observations that are good quality, you can have better predictions that can inform decision-making, policymaking and hopefully that will help look after the oceans and keep them healthy for everybody on this planet.

The AUV Sirius on the aft deck of the R/V Falkor. (SOI / Logan Mock-Bunting)

Oceans Deeply: What are the challenges in designing these autonomous vehicles to map the seafloor?

Pizarro: There are plenty of them – just ironing out glitches in software for everything from the logging instruments to making control decisions, to executing missions where you’re trying particular behaviors.

There’s a lot of logic built into it, in terms of how to behave when something goes wrong – what to do at the end of a mission when you haven’t heard from your human overlords. There’s a lot of smarts built into these systems. Some of that may not be fully tested, and so there’s a bit of iteration just getting all that stuff glued together and talking end-to-end. Sometimes, when you add a new instrument or a new capability, that may break or may need to be updated, or sometimes one of your instruments has an internal upgrade, and that produces other problems.

Then, hardware-wise, if it’s a new system, some of the things may not have spent enough time being run through the paces and confronted with the range of demands you may expect in terms of currents, or depth, or simultaneous load of thrusters going on while using high-power sampling interference.

Then there’s just operating in a new environment. You tend to find that when things are operating, the situation may change faster than you realize, and then these autonomous systems have to make these decisions on their own. Often they’re doing the right thing, but sometimes they’re not and that’s where you have to go in and iterate again. So, the challenges are also in learning from using these things in real-world situations.

An IVER AUV is moved via crane. (SOI/ Monika Naranjo Gonzalez)

Oceans Deeply: What are the benefits of having multiple drones that are doing the same thing?

Pizarro: The point of having multiple platforms is that you should be able to perform a survey more quickly, or collect more data for the same amount of time you’re spending at sea.

One of the greatest limiting factors is ship time, so it’s more about how do you run in parallel this type of work so that you’re multiplying the capabilities of a ship like Falkor.

And the challenges there are in this planning. “Oh, we’re supposed to be here at 10 a.m., and also 2km (1.2 miles) to the west at 10 a.m.” Can’t do both, right? So what are we going to do? The tools that help us deal with having many things in the air at the same time is one of the aspects that we’re also trying to develop further on this trip.

What we’re trying to get here is showing multiple vehicles operating together in a way that’s sensible in that we’re getting high-quality scientific data out of them without swamping the human operators and without requiring too much of an army of specialists to do so.

The Lagrangian Float weighs 60lb (30kg), making it much simpler to manage than most AUVs. (SOI/ Monika Naranjo Gonzalez)

Oceans Deeply: What does your work on this expedition tell you about the future of drones and humans in oceanographic research?

Pizarro: I’d like to think that the future is bright. These platforms and their capabilities are only going to get better. I think what really motivates me in the long term is how to make them more cost-effective and scaleable. So how do we deploy a large number of autonomous systems and sample the ocean at a much broader and intensive scale than we do today, without it costing hundreds or thousands of times more? We’re not going to spend much more money than we do on this. How do we get more out of the money we’re spending on it?

It seems that autonomous systems that ultimately don’t rely on having people in the field is the way to go. These demonstrations, these tools that we’re slowly putting together, will likely play a role in that type of scenario – where you have an autonomous surface vessel that deploys unmanned underwater vehicles away from civilization with light oversight from people, doing fairly complex operations, like imaging the sea floor. And it’s not costing you tens of thousands of dollars to just be there.

The University of Rhode Island’s Lagrangian Float is small, light and easy to deploy. (SOI/ Monika Naranjo Gonzalez)

Suggest your story or issue.


Share Your Story.

Have a story idea? Interested in adding your voice to our growing community?

Learn more