Random Walk
E/Q Phone Home
Pity the poor news anchor in the immediate aftermath of an earthquake. While he’s waiting on the feed from the Caltech Seismo Lab, there’s precious little for him to do, apart from restating the obvious (“If you’ve just joined us, there was an earthquake moments ago”), offering unassailable prognostications (“Aftershocks are certainly a possibility”), and taking breathless phone calls from chatty viewers (“Nah, this one was definitely stronger than Northridge, because this time both my cats freaked out”).
The banalities and generalities will no doubt be with us forever. But that thoroughly unscientific community survey may soon change into something much more useful.
What if there were inexpensive, pocket-sized gizmos that could sense the shaking and zap that info to Seismo while the ground was still rolling? What if you could deploy a million such devices throughout Los Angeles, forming an incredibly dense sensor network?
Oh, wait—this gizmo already exists. It’s your cell phone. There’s an accel-erometer built right into it; add a bit of software, and you’ve got a portable seismometer. Instead of a grainy image of a major earthquake collected from widely scattered recording stations, seismo-smart phones could give first responders a high-definition, block-by-block (or even floor-by-floor) picture of how buildings and soil are moving.
So says Ramo Professor and Professor of Computer Science K. Mani Chandy. He’s helping develop the Community Seismic Network (CSN), a proposed city-scale system that would link low-cost sensors with a cloud-computing network. The CSN’s detailed maps would allow emergency crews to be dispatched where they’re most needed when they’re most needed immediately after the quake. Says Rob Clayton, professor of geophysics, “In a quake,one building may be destroyed while the building standing next to it is not damaged at all. We saw examples of this in the Mexico City quake. The CSN enables the community to build a highly detailed shake map collaboratively, giving first responders much more accurate information.” Clayton is the principal investigator on a grant from the Gordon and Betty Moore Foundation to deploy a prototype network of some 1,500 sensors throughout Pasadena.
In between earthquakes, the CSN has other uses. Tom Heaton (PhD ’78), professor of engineering seismology and another PI, proposes to put dense sensor networks into high-rise buildings, with many sensors on each floor. “Being able to monitor a building’s behavior over long periods of time will help civil engineers better understand its dynamics,” he says. “Add to that a dense network of ground sensors, and seismologists and civil engineers can work together on unraveling the interaction between ground soil mechanics and building mechanics.”
“It’s a nontraditional way of doing seismology and civil engineering,” acknowledges Monica Kohler (PhD ’95), a senior research fellow in mechanical and civil engineering. Earlier this year, she and a group of students did a preliminary test in the nine-story Millikan Library, the tallest building on campus. The students distributed a dozen smart phones throughout the building and on the roof. A Caltech staffer videoing the event for posterity contributed a Caltech beaver-mascot bobblehead and a cup of water to the rooftop sensor array, and then the shaking machine, an off-axis rotary contraption that has long been a staple of Caltech’s earthquake engineering studies, was loaded with weights and set spinning. Small ripples immediately appeared in the water.
Graduate student Ming Hei Cheng (MS ’09) explains: “Although there’s only about 500 pounds of weight in the shaker, the rotation makes it act as if it were 5,000 pounds. And here on the roof is where we’ll feel the maximum acceleration and the maximum displacement.” In fact, the entire nine-story building was moving as much as a couple of inches from side to side—easy enough for the sensors to detect. (The bobblehead remained motionless.)
The CSN is an integral part of a year-long course on distributed systems being cotaught by Chandy and Julian Bunn, a lecturer in computing and applied mathematics, along with grad students Michael Olson and Matthew Faulkner. Besides the Moore Foundation, the project is being supported by the National Science Foundation’s Cyber-Physical Systems Initiative, and by Google, which donated 20 Android phones. The other CSN team members include project manager Richard Guy, computational scientist Leif Strand, grad student Annie Liu (MS ’10), undergrads Rishi Chandy and Jonathan Krause, and Andreas Krause, assistant professor of computer science and the third PI. Says Chandy, “The project is typical of Caltech’s multidisciplinary approach to research, with undergraduates, PhD students, staff, and faculty from different areas collaborating to solve real problems that impact humanity.”
But the CSN won’t rely solely on smart phones. The experiment on the Millikan rooftop also tested several freestanding miniature sensors, each the size of a half-dollar, that can be plugged into a desktop computer. Arming a PC with one of these devices, at a cost of about $100 each, makes it possible to record several months’ worth of motion data. “Everybody’s got a computer,” remarks Kohler. “So we could distribute hundreds of thousands of these. And in ten years, everybody will have a smart phone.”
So right after that next big temblor, when a new crack has appeared in your driveway and you’ve just got to tell everyone, don’t be in such a hurry to grab that cell phone. It may have already beaten you to it. —DZ
Monica Kohler (in the yellow shirt) supervises the deployment of smart phones on the roof of Millikan Library. Also in the photo: grad student Shiyan Song (MS ’09).
