|
Wireless
Sensor Networks Spread to New Territory New York Times, July 26, 2004 -- For years it was mostly the stuff of science fiction. But now communications technology has advanced to the point where sensors, machines and computers are beginning to talk to each other wirelessly in ways that could dwarf traffic between familiar devices like cellphones. Wireless networks of sensors and machinery have long been used in a handful of critical tasks like air traffic control. Now technology is expanding their reach into areas like managing lighting systems and detecting construction defects. But that is just the beginning. As an example, a wireless communications system is being tested to tell precisely when to irrigate and harvest grapes to produce premium wine. Other systems under development monitor stresses on aging bridges to help states decide maintenance priorities, and warn when oil tanker equipment is in danger of failing. "The range of potential market applications is a function of how many beers you've had," said Tom Reidel, co-founder and senior vice president for business development at Millennial Net, a start-up in Cambridge, Mass. Hans Mulder, associate director for research at Intel, which has invested heavily in developing chips for wireless communications, says wireless chatter among sensors and machines "will be pervasive in 20 years." By then, he said, companies will be producing hundreds of devices to support the communication of machinery through the ether. Futuristic though it may seem, the movement has captured the interest of venture capitalists, whose investments are supporting start-up sensor networking companies like Millennial Net, Crossbow Technology and Ember. And for every company already involved, several others are sizing up their opportunities. "As a revolution it's not here yet, but as an idea it's everywhere," said Robert A. Birdsey, a market development manager for Dresser Instruments, a subsidiary of Dresser in Stratford, Conn., that makes a range of pressure and temperature gauges. Mr. Birdsey made his comment during a break at a recent trade show in Detroit, where he was researching whether Dresser needs to add wireless communications capabilities to its products. The grandest visions include concepts that still sound far-fetched, like "smart dust" - a term used to describe communicating sensors no bigger than grains of sand that might be strewn by the thousands on fields and forests. Smart-dust proponents envision such sensors being used to monitor forests for fire, warn soldiers of dangerous substances on the battlefield and alert border guards to activity in remote areas. No one has yet come close to making components small enough for smart-dust systems. But a concept known as mesh networking has reduced the power requirements, at least theoretically, to the point where researchers expect to produce, within the next few years, networks of sensor nodes the size of postage stamps powered by coin-size lithium batteries. In such networks, sensors would need only enough power to communicate with their neighbors. Messages would be passed along to a more powerful control station or central computer. All this has whetted the interest of investors looking for the next big thing in technology. A market research firm in San Diego, ON World, estimated that wireless sensor networks generated less than $150 million in sales last year but will top $7 billion by 2010. Wireless sensor networks can be deployed in hours in places where it could take months to set up wired networks. Analysts estimate that wiring accounts for 80 percent of the cost of many sensor networks. And wireless networks, unlike wired systems, can be easily rearranged: moving a node can be as simple as pulling it off a Velcro mount and sticking it somewhere else. Each point or node on the wireless network needs components like radio chips and an antenna, and each node also needs a battery or some other power source. But individual nodes, already sometimes as small as a deck of cards, are steadily shrinking as more functions are integrated into their digital cores. The nodes broadcast data to a central computer for processing, either directly or, in large networks, through intermediate traffic management devices called gateways. Wireless networks
could keep a close eye on situations that are impractical to monitor
now because of the high cost of human inspectors or wired systems. The nation's largest processor of sugar beets, for example, has been working with Pedigree Technologies, a start-up in Oxbow, N.D., to develop sensor networks to monitor conditions in the beets that it must store after harvest for up to eight months in 20-foot-high piles. The goal is to detect pockets of heat buildup before the rising temperatures start to break down the sugar in the beets, said David Berg, vice president for operations at the company, American Crystal Sugar. Sensor networks are being designed not only to collect data but also to serve as the nervous system for electronic controls that turn other machines on and off in response to changing conditions. For example, the Defense Advanced Research Projects Agency is using technology developed by Sensoria, a start-up in San Diego, to create mobile antitank land mines that automatically shift position to fill gaps after other mines have detonated. As the industry looks forward, "reliability, power and cost are the big issues,'' said Kristofer Pister, the chief technology officer at Dust Networks in Berkeley, Calif. Dr. Pister, who coined the term smart dust as a researcher at the University of California, Berkeley, said one unfortunate consequence of the catchy phrase had been to focus too much attention on efforts to shrink the nodes. Indeed, powering the sensor networks remains a far more pressing challenge in many potential applications. Power can be conserved by leaving sensors in sleep mode most of the time; internal timers or network signals activate them only when they are needed to take a measurement or pass on a signal. So far, though, most mesh systems have proved far more error-prone in the field than in research settings, underscoring the need for hardware and software tools different from those used in the consumer wireless sector. Unlike people, machines can exchange data about anything - temperature, moisture levels, pressure, vibration and so on - in small fractions of a second. Thus, they rarely need the huge swaths of bandwidth people require for wireless communications. Signal interference and security are of bigger concern. While wireless networks for people assume that, when the connection fades, users will move to find a stronger signal, designers of sensor technology have to account for the fact that the machines cannot move themselves. Some of the issues were addressed by a committee of the Institute of Electronic and Electrical Engineers, which developed a new design standard for low-power wireless technology. Meanwhile, an alliance of companies led by semiconductor giants like Motorola, Philips and Samsung and pioneering companies like Ember has been working on supplemental standards to add networking and security specifications. "There's a whole ecosystem of hardware, software and service guys springing up," said Mr. Reidel of Millennial Net. Media Contact: Back
to News
|
