DTN Networking to Monitor Ecosytems

Katie Huston for TEI

Computer scientists often sit behind their desks, but Associate Professors Brian Levine and Mark Corner in the deparment of Computer Science are getting out of their offices and into nature.

They’re applying their work with Disruption Tolerant Networking (DTN) to environmental problems, including monitoring underwater environments and tracking wildlife.

Although their work is driven by the needs of scientists, Levine says, it’s also a rewarding technological challenge.

“It’s not an application of known ideas in computer science. We really had to push the envelope to develop these kinds of networks,” he says.
Unlike the Internet, which assumes constant connections, DTN is designed to allow data exchange in environments where users or computers are disconnected most of the time. The technology is able to store data constantly in intermediaries, and transmit it intermittently when a connection is available.

“It’s a lot like the Pony Express,” Levine says. “You get this brief period of connectivity, and you keep going.”
Levine and Corner began using the technology on PVTA buses in May 2004. Two years later, Levine says, “It was quite obvious that what we were doing terrestrially could be used underwater.”
With current undersea monitoring methods, cameras and sensors are planted in the water, connected to shore by an expensive fiber optic cable.

However, existing technology doesn’t work well when monitoring large areas, like the Quabbin Reservoir, the principal source of drinking water for eastern Massachusetts – and it’s also prohibitively expensive.
“It covers a huge area,” Levine says of Quabbin. “Nobody has the money to monitor something that large… To do that with the existing technology, there would have to be many of these observatories very densely located next to each other wired together with fiber optics.”

Using DTN underwater, though, allows access to a larger area, makes monitoring more affordable, and makes information more accessible and easier to disseminate.
Levine’s approach removes the need for a connected network. Instead, information is relayed through acoustic networking, rather than radio frequency, which does not travel underwater, and fiber optic connections. Levine’s design also makes use of floating buoys, which can pick up the information and relay it to passing boats or onshore monitoring stations.

“There are no wires, which makes it flexible, and it allows you to deploy equipment sparsely, which means it’s cheaper,” Levine says.

In October 2007, he carried out a test deployment in Quabbin Reservoir, and the results were promising. He hopes that a DTN network could be running in the Quabbin next summer, to help monitor water quality and animal interactions with the environment.

“If we can advance our technology, it’ll really change the way scientists operate. That kind of synergy really gives us a driving motivation to discover new technologies,” Levine says.
Mark Corner, who’s been at UMass since 2003, is applying DTN technology to another environmental problem: tracking turtles.

Right now, tracking turtles is pretty labor-intensive, Corner says. Scientists attach small radio transmitters to the backs of turtles, then go out in the woods and use a directional antenna to locate turtles. Sometimes, it can take a few hours to locate one animal.

“That’s pretty labor intensive, and you’ll only get a GPS estimate for each turtle maybe once a week, once every few days,” Corner says.

He’s trying to improve the system by putting a small wireless computer, about two by three inches, on the turtle’s back with a GPS receiver, a wireless radio, flash memory storage, a battery and a solar cell.
The units collect GPS readings on a regular basis, store the information locally, and transfer the data to other turtles or to a base station. The base station also has a cellular modem, so it’s capable of emailing data when in range of cellular service.

It hasn’t been easy. “Computer science that you do in the comfort of your office tends to be a little more predictable,” Corner says. “I had to worry about things I don’t know anything about, like glue and acrylic and waterproofing things.”

It’s also a challenge to make the units reliable. “When your computer crashes, you turn it off and back on. The turtle is not going to do that for you,” Corner says. “It has to run for months at a time.”

Although he originally hoped to track wood turtles, a threatened species native to Western Massachusetts, the current technology is too large to use on wood turtles, which are usually between five and nine inches long, so Corner is looking at removing the GPS unit to make the computer smaller and lighter.

“If you make things half as big as they were, you can greatly increase the number of species that you can monitor,” he says.

For now, Corner and Mike Jones, a Ph.D. student in Wildlife and Fisheries Conservation and Organismic and Evolutionary Biology, are trying the units on snapping turtles. He also hopes to use DTN technology to monitor other turtles.

Despite the challenges, Corner and Levine are excited about DTN technology’s potential. “There are a lot of different applications here, and if you can make these things small and cheap and fairly easy to deploy, you can start monitoring multiple species at the same time. You just sort of get this massive explosion of data,” Corner says. “Scientists always want more data.”

Right now, environmental monitoring relies largely on graduate students going out to collect data. However, DTN may facilitate a more continuous stream of data, freeing up graduate students to analyze the information. It may also allow longitudinal studies, enhancing scientists’ understanding of environmental trends.

“Our quality of life is really based on the quality of the environment,” Levine says. “It is critical that we understand a model of how our actions improve or diminish the ecological cycles in play around us.”

Top of Page