A Nest of Sensors
by David Pescovitz
 Printer-friendly
version
S. Shankar Sastry is the NEC Distinguished Professor of Engineering (Peg Skorpinski photo)
|
On a sunny August day at UC Berkeley's secluded Richmond Field Station, several graduate students are erratically running back and forth through an overgrown field. Meanwhile, small unmanned aerial vehicles buzz overhead. The students are playing a game but it's not football, soccer, or some other college sport. They're playing Multiple Target Tracking and Pursuit Evasion Games, and it's serious business. The engineering students and faculty are demonstrating a massive wireless sensor network that's the end result of a multi-year Defense Advanced Research Projects Agency research contract.
"We think that the infrastructure of tomorrow involves large, self-configuring wireless sensor networks for everything from environmental monitoring to the control and security of electric power, water, petroleum, and natural gas systems," says principal investigator Shankar Sastry, director of the Center for Information Technology Research in the Interest of Society (CITRIS).
Launched in 2001, the aim of the Network Embedded Systems Technology (NEST) project was to build an experimental platform to speed the development of sensor network technology. The project is a collaboration between Sastry and electrical engineering and computer sciences professors David Culler, Eric Brewer, Kris Pister, and David Wagner. The research team also included graduate students Songhwai Oh, Phoebus Chen, Shawn Shaffert, Bruno Sinopoli, Jaein Jong, Sukun Kim, Prabal Dutta, Kamin Whitehouse, Gilman Tolle, Jonathan Hui, Jay Taneja, Tanya Roosta, and Bonnie Zhu. Research staff members Mike Manzo and Cory Sharp contributed too.
An "evader" traverses the field of sensors. (Songhwai Oh photo)
|
At the heart of NEST are "motes," wireless sensors invented at UC Berkeley that keep a constant vigil on temperature, light, motion, and myriad other factors. Once deployed, the motes self-organize into networks that pass along data, bucket-brigade style, in short hops until the information reaches a central computer for processing. If a mote breaks down, the network routes around it.
In recent years, UC Berkeley engineers have experimented with a wide variety sensor network applications, from smart energy monitoring in homes to detecting the seismic stability of buildings to studying an elusive species of seabirds in their natural habitat. According to Sastry though, the only way the motes will move into mainstream use is if they're "robust, easy to program, and simple to deploy."
The network tracks two individuals moving through the field of sensors. (courtesy the researchers) [view larger image]
|
"We set up the NEST demonstration so the stakeholders, the people who would actually use sensor networks, could assess whether the technology really works," he says.
The NEST demonstration involved 573 motes outfitted with passive infrared detectors, microphones, and magnetometers. Distributed over two square kilometers outside, the sensors were solar-powered, remotely programmable, and resistant to most of what nature could throw at them.
"We had to install little protrusions on the motes to keep birds from sitting on the solar cells," Sastry says.
One of the sensor nodes mounted on a tripod. (Songhwai Oh photo)
|
In the end, the winner of the Multiple Target Tracking and Pursuit Evasion Games was the network itself. As three graduate students moved through the field of motes, their paths were tracked in real time. Meanwhile, on the screen, simulated "pursuers" stayed hot on their trails. In the real world, the military might use such a network to track enemy forces. Pursuers could be people or robots, Sastry says.
As the games were played, small remote-controlled helicopters and fixed-wing unmanned aerial vehicles (UAVs) buzzed over head, collecting the data from the sensor network. Even zipping by at 55 miles per hour, the UAVs were able to
receive the wireless transmissions from the motes.
"In this case, the demonstration was all about chasing people," Sastry says. 'But there is also talk about oil refineries, water supplies, and other operations using sensor networks instead of their existing supervisor control and data acquisition (SCADA) wired systems."
Quickly identifying problems in large-scale infrastructures can be quite challenging, Sastry explains, requiring complicated and expensive hardware. Distributed sensor networks may be a more secure and cost-effective solution for fault detection and diagnosis, he says.
Now that the NEST network has proven itself, Sastry is reluctant to disassemble the system. He's now planning to keep the network alive as a true test bed.
"I want people to do science rather than demos," he says. "How easy is it to reprogram the motes remotely? How will they do out in the field over long periods. We might get some rain showers soon."
Shankar Sastry's home page
NEST Final Experiment, 2005
NEST Project
Berkeley Wireless Embedded Systems (WEBS)
Center for Information Technology Research in the Interest of Society (CITRIS)
Lab Notes is
published online by the Marketing and Communications Office of the UC Berkeley
College of Engineering. The Lab Notes mission is to illuminate groundbreaking
research underway today at the College of Engineering that will
dramatically change our lives tomorrow.
Media contact: Teresa
Moore, Lab Notes editor, Director of Marketing and Communications
Writer, Researcher: David
Pescovitz
Web Manager: Michele
Foley
Subscribe or send comments to the Engineering Marketing and Communications Office: lab-notes@coe.berkeley.edu.
© 2005 UC Regents.
Updated 11/1/05.
|
