The big idea: If you’ve ever played air guitar along with your old Beatles or Jimi Hendrix albums and imagined what your wicked licks would sound like, Kris Pister, M.S. ’89, Ph.D. ’92, has great news for you.
Over the past 15 years Pister, a professor of electrical engineering and computer sciences, has taken the theory of microscopic wireless sensors and made it a tiny reality with vast potential. He has figured out how to build millimeter-sized computers that can detect anything from light and heat to temperature and vibrations, and send that information on. The sensors—collectively, “smart dust”—have endless potential. They could monitor the humidity and temperature of foodstuffs on their way to the supermarket, or of each individual office in a 50-story building, increasing comfort and saving power. They could be placed throughout African river valleys to study disease-carrying tsetse flies. Winemakers could spread smart dust throughout their vines to detect stress or mold; bridge makers could monitor vibrations causing wear and tear; soldiers could place sensors along roads to detect insurgents’ IEDs.
And for fun, you could take some smart dust motes wirelessly linked to your computer, glue one to each fingernail, and play your very own air guitar.
After more than a decade of invention and frustrating delays in getting his technology to market, Pister says smart dust’s time has finally come. As a science-fiction fan, he’d been reading about such technology since he was a kid. So when scientists got serious about the theory in 1992, he saw the opportunity to make the fiction into fact. The mote-sized sensors he designed and christened had four parts: “sensing, computation, communication, and power.” Year by year, each of those components has become smaller, faster, and cheaper. The biggest challenge was creating electronic circuits that could run on “outrageously low power.”
At last, he says, “we’re on the verge of breaking through commercially,” thanks to an unexpected application now in use just up the road at the Chevron refinery in Richmond. Pister explains that the refinery is home to about a million sensors, working to anticipate breakdowns and improve mechanical efficiency. Yet only 200,000 of them are wired up. Why?
“They estimate the average cost is over $10,000 per sensor to activate, because they have to run a kilometer or more of [fiber optic] wire to them.”
He can offer factory bosses reliable connections for a fraction of that cost. “Now we have this beachhead market, and that means the lines will increase, the costs will come down, and eventually we’ll be able to hit this dream of having just this very low cost sensing of everything, everywhere.”
What’s next: Pister admits he’s unlikely to see sensors reach dust size in his lifetime. Companies are happy with the dime-sized to pager-sized sensors he can supply now. But his research continues, and in the next year or two he hopes to have sensors, including a solar panel and antenna, that are less than one-quarter the thickness of a piece of paper.
“That’s as small as I can picture it getting,” he says, “but something that floats in the air is pretty close to dusty, right?”