Seafaring robots can explore expanses of the ocean where data collection might otherwise prove costly, time-consuming and dangerous. Like any intrepid explorer, however, they can lose their way.
Scientists in the University of Delaware's College of Earth, Ocean, and Environment (CEOE) recently demonstrated their search and recovery prowess when an autonomous underwater vehicle (AUV) went missing in the Cayman Islands.
The equipment hid wedged along a submerged limestone wall for four days until they were able to use another robot - fittingly enough - to pull it up to the surface.
"Especially with uncharted areas, you never know exactly what you are going to run into," said Hunter Brown, CEOE's advanced underwater systems manager. "It's an operational hazard."
Brown and Matt Oliver, assistant professor of oceanography in the School of Marine Science and Policy, were testing the AUV off Little Cayman in January for use in a new study abroad program on coral reefs and scientific diving. Undergraduates helped assemble underwater robots, attended lectures on the technology and conducted dives that complemented scientific sampling done by the AUV.
The reef structures around Little Cayman have never been systematically surveyed, so the researchers sought to create bathymetric maps of seafloor contours and take digital photographs.
Starting off the island's southern shore, their first test to make sure the vehicle was working went fine. Next, they programmed the robot to record data along an hour-long course at a depth of roughly 60 feet.
"We pressed 'go,' and it never came back," Brown said.
Keeping tabs on an AUV when a mission is underway is not easy. The device can only transmit its GPS location via satellite when it comes up to the surface because it works on radio waves that don't penetrate water very well. A mission simulator gives an estimated location based on the planned path, but it's only just that: an estimate.
As the clock ticked past the AUV's scheduled surfacing time, the researchers thought the ocean current might have moved the robot too far from the boat to see it. Once it was 30 minutes late with no satellite phone message, however, they knew they probably had a problem - and nerves started to fray. It was the only AUV they had on the trip, and such equipment can easily be worth more than a Rolls-Royce.
The team suspected that the mission plan accidentally sent the AUV over a steep drop-off the coral reef and down a limestone wall. The charts they had for mission planning were out of date and the drop-off was not precisely indicated, Brown said. The AUV may have tried to turn around, but instead got stuck along the wall and surging water pushed it into a small cavity.
Later in the day they used an acoustic modem to make contact with the AUV, which broadcast signals before the battery eventually died. The messages indicated they were close, and they were able to estimate a depth of 550 feet - nearly 500 feet deeper than it was supposed to be.
For reference, recreational scuba divers only go to about 130 feet, and expert technical divers rarely go past 250 feet.
Brown had already alerted colleagues back on UD's campus, who started to mobilize to bring another type of a robot, called a remotely operated vehicle (ROV), to Little Cayman. Fortunately, UD has one of the larger fleets among seagoing universities with a total of six underwater vehicles.
Art Trembanis and Carter DuVal of the School of Marine Science and Policy came down to Little Cayman to help, and their frequent collaborator Val Schmidt of the University of New Hampshire helped calculate areas to search. Undergraduate students assisted in preparations for the search and recovery effort.
"It was a real team effort all around," Trembanis said.
The ROV is connected by a long cable to a ship and controlled by a joystick. Operators maneuver it while viewing the ocean through its camera, and the device has a short arm off the front for clamping down onto objects.
With the search area pinpointed and the team convened, they set out on the fourth day to recover the AUV. After about 40 minutes of searching along the limestone wall, DuVal, an oceanography graduate student, asked "Is that something?" as he operated the ROV and saw the faint outline of the missing robot on the screen.
From there, the question was whether the small arm would be able to pry the AUV out of the cave. The operation is comparable to arcade claw machines where a joystick is used to try to grab a stuffed animal or toy - but with limited visibility, strong ocean currents and slippery conditions.
"I made a go at it, and I was lucky that I hit the bar on that first grab," DuVal said.
With the ROV's cable stretched nearly to the limit, DuVal carefully teased the AUV off the wall at a depth of 506 feet and brought it to the surface. With a buoyancy difference between the two robots, the recovery was tricky all the way up.
Eventually, they brought it safely on board with only minor scratches to the plastic parts. The four-day recovery showed that knowing how to retrieve a missing robot is just as important as knowing how send it on a mission in the first place.
The team shared their story in the March 2014 issue of Ocean News and Technology to help other underwater robotics user that might face similar situations in the future. Still, there are no guarantees when it comes to ocean research.
Trembanis points out that the field of oceanography has a legacy of managing risks against the rewards, ever since the first worldwide ocean research expedition set sail on the HMS Challenger with no GPS or weather forecasts in the late 1800s.
"People have always been pushing the technological boundaries to deal with this under-sampled ocean that we have," Trembanis said. "But there's a saying in oceanography: 'Don't put it over the side unless you're prepared to lose it.'"