On the Beach with the Gaviators
Two Gavia autonomous underwater vehicles side by side on the shore of Lake Pavilion. There are only twelve of these AUVs in the world so it’s rare for two of them to be seen together. Credit: Henry Bortman
June 30, 2010-
Today I hung out on a pebbly beach with a group of scientists everyone refers to as The Gaviators. They’re named for the Gavia AUVs (autonomous underwater vehicles) they work with. A Gavia AUV looks like a brass-colored torpedo, about 6 feet long and 10 inches in diameter, with a pointed nose at one end and a propeller at the other. In-between are various data-collection modules that can be configured as needed.
Gavia is good for large-scale, comprehensive, repetitive work. It doesn’t get tired, hungry, bored or distracted, and, unlike a DeepWorker pilot, it flies in straight lines. Once it’s programmed, it can run for hours, traversing back and forth across the lake in a “lawnmower” pattern, collecting vast amounts of data without anyone having to worry much about it. When it works.
A Gavia unit owned by the University of British Columbia (UBC) has been part of the summertime work at Pavilion Lake for several years. Andrew Hamilton, a Ph.D. student, and Claudine Fortier, who recently completed her undergraduate studies, were operating the UBC unit today. Alex Forrest, another Ph.D. student also on the UBC team, doubles as a DeepWorker pilot so he was off flying a sub.
Shortly before the PLRP field season UBC’s Gavia got a software upgrade. It crashed the whole vehicle, and resolving the problem required new hardware. That caused more software problems. Anyone who’s ever upgraded Windows or Mac OS knows is familiar with this story. Ultimately, the team had to summon Richard Yeo, a Gavia expert, out from Iceland to troubleshoot. So UBC’s Gavia had a slight handicap from the outset.
Claudine Fortier and Andrew Hamilton of the University of British Columbia reassemble their Gavia lakeside in preparation for a data-collection flight. Credit: Henry Bortman
Just to make things more complicated, this year for the first time the UBC Gavia team was joined by a second team from the University of Delaware (UD), who brought a second Gavia to the lake. The UD team is led by Assistant Professor Arthur Trembanis; working with him are Ph.D. student Stephanie Nebel and Masters student Jonathan Gutsche. The two teams are hoping, before the time this year’s field season ends, to be able to deploy both Gavias simultaneously, working in tandem with the two DeepWorker submersibles.
This collaboration has two goals. One is to study the best way for the robotic Gavias and the human-piloted DeepWorkers to operate together. If there are ever to be human missions to the moon or Mars, the humans will undoubtedly be accompanied by robotic assistants, so this is a good opportunity to practice.
Stephanie Nebel of the University of Delaware and Richard Yeo, a Gavia consultant, haul Gavia out of the water after checking to make sure it is seaworthy. Credit: Henry Bortman
The other, more-immediate goal is to explore a salt-water “lens,” a layer of water with a relatively high concentration of dissolved calcium carbonate and other salts, present in the bottom 10 meters of the lake. The scientists know the salinity of the lens varies over time; they want to better understand these changes as well as variations in its size and shape. The microbialites, you may recall, are calcium carbonate formations, so the dynamics of the salt-water lens may hold a chemical clue to microbialite growth patterns.
The plan was for the DeepWorkers to deploy CTD sensors, which measure salinity and other water characteristics, in the salt lens along the bottom of the lake, where it can be difficult for Gavia to maneuver safely. Meanwhile, the Gavias, also outfitted with CTDs, would fly some 10 meters above the DeepWorkers, collecting data from the water overlying the lens.
That was the plan. But the collaborative effort hit a few snags. The people get along fine. It’s the equipment that’s been uncooperative.
Gavia uses an acoustic modem to report its status when it’s underwater. So do the DeepWorker pilots, when they talk to the nav boats. (See my June 29 blog entry.) And the two modems use similar frequencies. So the first time Gavia came near the DeepWorkers, it fouled their communication. That’s not so good when you’re flying around in a tiny submarine 150 feet below the surface, and the only way you can let the people up top know how you’re doing is to talk to them.
So Gavia’s acoustic modem had to be turned off. Then Gavia-with-silent-modem had to be tested to make sure the modem was no longer causing a problem for the DeepWorkers.
Hamilton, Yeo and Fortier check Gavia’s programming before sending it off on a mission to measure the lake’s water chemistry. Credit: Henry Bortman
It wasn’t. But there’s more. To get highly accurate positioning the UBC Gavia uses a sonar system to navigate underwater. Yesterday, however, the UBC team found that the Gavia sonar system and the system used to track the position of the DeepWorkers interfere with each other. So the Gaviators had to take both AUVs apart, and the UBC team had to borrow a module from the UD team. The module is a navigation system that uses lasers and gyroscopes instead of sonar and doesn’t scramble the DeepWorkers’ navigation. Problem solved, but one less AUV in the water.
When all that was done, mid-way through the afternoon, a single hybrid Gavia was ready to do some science. The DeepWorkers were lowered into the water, and Gavia was sent off on what was supposed to be a 3-hour mission of trouble-free exploration.
It did manage to get some useful science done, flying back and forth across the lake a couple of times collecting CTD data. But after about an hour, Gavia popped to the surface and phoned home to report that it had aborted its mission. That problem has been identified and the team is busy making some minor repairs.
Meanwhile, the University of Delaware team plans to produce a high-resolution bathymetric (depth) map of the entire lake using a type of imaging sonar known as GeoSwath. A great deal of sonar data is already available from past work at the lake, including a partial low-resolution bathymetric map.
A detailed bathymetric image of a rockslide along the shore of Pavilion Lake, produced by data collected by the University of Delaware’s Gavia AUV. Credit: Art Tremblanis / Richard Yeo
By overlaying information about microbialite shapes and sizes, temperature, depth, water chemistry and other data on this high-resolution map of the lakebed topography, researchers hope to be able to visualize large-scale trends that may explain why certain types of microbialites form where they do.
The UD team has managed to complete a couple of very successful GeoSwath runs so far, including one this morning that mapped a shallow area between the central and north basins filled with large mounds. They were hoping to continue their work Wednesday afternoon, but then they had to take their Gavia apart so the UBC Gavia could conduct its experiment with the DeepWorkers. The GeoSwath project will resume soon. But first the UD Gaviators have to get their navigation unit back from their UBC colleagues.
Don’t worry too much about the Gaviators’ technical difficulties — they’re really smart people and quite skilled at troubleshooting.