Smart Dust Comes Out of the Labs

Smart Dust Comes Out of the Labs


check out the article here...

Imagine an Internet of things, where everyday objects, rooms, and machines are connected to one another and to the larger digital world. Sensors on expensive factory equipment would tell you when the machinery is about to fail. Cargo shipping containers could search their contents for nuclear material or other hazards. Every office in a high-rise could report its temperature and humidity and whether its lights are on or off. Each foot of the nation's streets and highways could monitor traffic flow, and every acre of farmland could keep tabs on the moisture level in its soil.

Kris Pister's eyes shine when he talks about such possibilities and about the so-called smart dust -- wireless sensor motes -- he's been developing since the late 1990s that will make them happen. "The idea," he says, "is connecting sensors in the physical world to data systems." The trick, though, is to do it wirelessly and with very low power. His smart dust technology was originally funded by the Defense Advanced Research Projects Agency, back when he was an engineering professor at the University of California at Berkeley. In 2002 he founded Dust Networks [www.dustnetworks.com] to commercialize the technology. On Monday he will finally announce that he has a working product ready to be shipped.


"Within 50 feet of where we are sitting," he notes, looking around the conference room near my office on the 17th floor of the Time & Life Building in New York City, "there might be a few dozen sensors: temperature sensors, air flow sensors, pressure sensors, vibration sensors on motors, humidity sensors, gas sensors in the HVAC system, even occupancy and light sensors." While each of those sensors might cost only $10 or $20, it can cost hundreds of dollars to wire up each one. Making the sensors wireless means you could place them anywhere you want, monitor more things, and pay less to do so. For owners of commercial real estate or even retail stores, that could offer a means to use less energy.


Sensors may seem mundane, but they are a $42 billion industry, according to Dust Networks. But the company isn't building sensors. Rather, it is building tiny radios the size of a microchip that can link sensors together into a wireless, peer-to-peer network. "Think of this as sort of like a NIC card for sensors," says CEO Joy Weiss, referring to the network interface cards that link computers to data networks. In this case, the sensors are creating their own wireless network by linking to other sensors within range. The data hops from sensor to sensor, just like on the Internet, except that Pister had to come up with a lighter-weight communications protocol than TCP/IP, which is the industry standard. The other difference, Pister says, "is if you leave your radio on all the time, your battery will be dead in a matter of weeks." To solve this problem, he figured out a way to turn down the radios almost all of the time, so they're just faintly listening for signals. They eat up so little power, he adds, that they'll last for three years on AA batteries.


More important, Pister says he's overcome the biggest barrier to the commercialization of this technology, namely the ability to scale up the wireless networks beyond a handful of motes. "Customers do not want a network where everything looks fine until you hit a certain point, and then you add three sensors and the whole thing breaks down," he says. His first networks will be designed to tie together 20 to 50 motes, and can handle "networks much larger than that," he says. But before smart dust can truly take off, networks of hundreds or even thousands of separate motes need to operate without a hitch.


If Pister hasn't already figured this out, somebody else will, as he is not alone in his quest. Competitors include other startups such as Ember [www.ember.com] and Crossbow Technology [www.xbow.com]. As smart dust becomes industrial-strength and more widespread, we'll start seeing companies slapping sensors on every piece of machinery and in every office so that they can measure things like their manufacturing processes and energy consumption in real time. That should lead to businesses that are run even more efficiently than they are today, from better security systems (think vibration sensors at property perimeters) to more productive farming (imagine a vineyard putting sensors in its soil to constantly test its moisture and alkaline content). When smart dust is everywhere, the physical world will be generating typhoons of data, and the new opportunities for mining and managing that data will be immeasurable.

g//