Tweeting from the Depths

USF’s solar-powered, underwater robot is able to analyze ocean conditions, report back to humans on land.


Video: Katy Hennig | USF News


By Vickie Chachere

USF News


ST. PETERSBURG, Fla. (March 27, 2012) – For years, University of South Florida marine scientists have relied on bright yellow underwater vehicles to patrol the Gulf of Mexico and report back to them important information on water conditions.


But none of these space-age vehicles have made quite the splash as the newest member of the team: @Tavros02.


Not only can this robot independently analyze what it is finding in the water’s depths, it is capable of using Twitter to report back what it’s learning about the waters it patrols.


@Tavros02 joined the Twitterverse on March 22 with a message that began: “My human says … “


Last weekend, the vehicle took over its own messaging as it traveled through the waters of Tampa Bay, sending information on its coordinates, the temperature and depth of the water it was patrolling.


David Fries, Director of USF Marine Ecosystems Technology Group. Photo by Vickie Chachere.

David Fries, director of USF's Marine Systems Technology Group, worked with a team at the College of Marine Science to make the high-tech breakthrough, turning a new generation of autonomous underwater vehicle into a unique marine observing and reporting system.


Unlike other underwater robot programs which wirelessly report ocean conditions back to researchers on land, the solar-powered Tavros vehicle can independently analyze water samples, compile the data and then convert that information into a message available to all who follow them on Twitter. Fries’ team has built a series of the vehicles, but for now just one is in the water busily Tweeting.


No humans needed for these social-media savvy robots to communicate what they know. “Twitter is their voice,” Fries said.


“This is a really exciting area of machine intelligence,” Fries said. “We were interested in getting machines taking in data and – like humans have processed that data in the past – Tweeting it back to us intelligently. We wanted to figure out how we could have them objectively communicate data without human input and using Twitter linked to artificial intelligent programs allows us to do that and distribute to a large user group.”


Fries has spent more than a decade inventing, patenting and improving on underwater sensing technology. In the months following the 2010 Deepwater Horizon oil spill, he was awarded a National Science Foundation grant to take the technology further than it’s ever been and created an underwater vehicle mapping system that could be more independent from humans than other iterations.


Three things make the Tavros (Greek for bull in a nod to the USF mascot and an acronym for The Autonomous Vehicle and Remotely-Operated Sensing) robots special:


Brian Gregson, biotechnologist at the College of Marine Science. Photo by Vickie Chachere.

Unlike previous generations of underwater vehicles, it is powered by solar cells instead of batteries which need to be recharged regularly. Second, the robots are equipped with devices which can analyze water on the spot and wirelessly transmit that data back to scientists, a process which at many marine science facilities has been done in the lab. And third, the robots have the ability to report that data back to land in a succinct, 140-character Tweet that can help scientists interpret what is happening in the environment in real time.


Autonomous underwater vehicles can be programmed to follow a specific route and dive thousands of feet under the water to test conditions at various depths. Capable of detecting such important markers as water temperature, currents, salinity, dissolved oxygen or microscopic algae and droplets of oil, the robots can be used for either routine surveillance of the oceans health or for specific environmental challenges and even monitor ports for security.


In 2010, USF’s gliders were key to detecting two vast underwater plumes that formed in the Deepwater Horizon oil spill, providing scientific evidence that the oil contamination of the Gulf was far more extensive than large surface slicks. More routinely, the gliders are used to help monitor for Red Tide or studying marine acoustics.


Gary Hendrick, a mechanical engineer who has been part of the Tavros project, said the key to taking the robots communications to the social media level was to build in a programming bridge which converts key data people would be interested in – such as water temperature and salinity – and convert that to written language.


For now, the robots are able to report on the data they collect, but the team is working on processing the data within the vehicle and translating that data into phrases that address what people really want to know: Are the conditions good for fishing? Is there a red tide bloom developing? Is their oil in the water, or is something suspicious happening underwater?


Brian Gregson, a biotechnologist with the Tavros group, said that capability is not far off. The team is working with David Millie at Florida Institute of Oceanography to create a set of standard messages linked to artificial neural networks which would interpret the data from certain water conditions the robots encounter, so that anyone – fishermen, beach goers and those curious about the health of the Gulf – can learn in plain English what the robots are sensing in real-time reports.


Standing with Tavros 5, one of the fleet of solar-powered autonomous underwater vehicles they have developed. In the water is Tavros 2, who has the capability to use Twitter and can be followed @tavros02. Photo by Vickie Chachere.

Increasingly, scientists are relying on autonomous surface and underwater vehicles to keep pace with the need for data on ocean health. Research cruises are costly in terms of time, fuel and supplies for the vessels, their crews and scientists. The data gathered is often limited in scope to the given day and place it was collected.


But robots can patrol continuously – USF’s battery-powered gliders can be at sea for up to a month or more. However when it’s time for their batteries to be recharged, the robots turn back and head for home – calling for a pickup when they are just off shore, a process that still requires a vessel, time and fuel to retrieve them.


Tavros is more independent than that. The vehicle can swim its pre-programmed route, resurfacing to recharge its batteries under the sun and transmit its data before continuing on its journey.


By coincidence, the off-the-shelf solar batteries that power the Tavros vehicles carry a familiar brand: BP.


Follow Tavros on Twitter: @tavros02