Palo Alto, California (CNN) — In the 1990s, a researcher named Kris Pister dreamed up a wild future in which people would sprinkle the Earth with countless tiny sensors, no larger than grains of rice.
These “smart dust” particles, as he called them, would monitor everything, acting like electronic nerve endings for the planet. Fitted with computing power, sensing equipment, wireless radios and long battery life, the smart dust would make observations and relay mountains of real-time data about people, cities and the natural environment.
Now, a version of Pister’s smart dust fantasy is starting to become reality.
The astronomer Carl Sagan once asked, “Who speaks for Earth?” Soon, the Earth may speak for itself.
That’s the goal of HP Labs Central Nervous System for the Earth, or CeNSE. The research and development program aims to build a planetwide sensing network using billions of tiny, cheap, tough and exquisitely sensitive detectors.
“We’re surrounded by technological assets that are deaf, blind, can’t taste, can’t smell and can’t feel,” says Stan Williams, an HP senior fellow who leads the Information and Quantum Systems Lab (IQSL). “CeNSE is all about giving all this compute power the awareness of what’s going on in the environment around it,” says Peter Hartwell, senior researcher and project team lead.
Hartwell envisions sensing nodes about the size of a pushpin stuck to bridges and buildings to warn of structural strains or weather conditions. They might be scattered along roadsides to monitor traffic, weather and road conditions. Embedded in everyday electronics, CeNSE nodes might track hospital equipment, sniff out pesticides and pathogens in food, or even “recognize” the person using them and adapt.
Taken together, that awareness could “revolutionize human interaction with the Earth as profoundly as the Internet has revolutionized personal and business interactions today,” Williams predicts.
For CeNSE to work, “we have to make sensors that are vastly more sensitive than anything else that have ever existed before, while being absolutely dirt cheap so that we can deploy them in very large numbers,” Williams says.
Hartwell is working on the first sensor to go into the field, a motion and vibration detector. More accurately called an accelerometer, Hartwell’s device is sensitive enough to “feel” a heartbeat. The source of that sensitivity is a 5mm-square, three-layer silicon chip. A portion of the center wafer is suspended between the two outer wafers by flexible silicon beams. When the chip moves, the suspended center lags behind due to its inertia. A measurement of that relative motion is used to calculate the speed, direction and distance the chip has moved.
This exquisitely sensitive accelerometer can detect a 10 femtometer change in the position of its center chip. That’s less than one-billionth the width of a human hair. As a result, it can measure changes to acceleration in the micro-gravity range. That’s about 1,000 times more sensitive than accelerometers used in a Wii, an iPhone or an automobile’s airbag system.
IQSL Lab researchers also plan to add sensors for light, temperature, barometric pressure, airflow and humidity.
While Hartwell’s accelerometer gives CeNSE its “feel,” the system’s “taste and smell” are just around the corner. Researchers in the group are using nanomaterials to boost a standard chemical and biological detection technology (Raman spectroscopy) to 100 million times its usual sensitivity rates. As sensitivity rises, sensor size can shrink. That could lead to detectors small enough to clip onto a mobile telephone. With a wave over produce, the sensor might warn consumers of salmonella on spinach leaves or pesticides present in “organic” produce, Hartwell says.
Strength in numbers
Today, sensitive detectors are expensive. The few out there tend to return a lot of false alarms. Deploying sensors en masse helps sort out background noise and hone in on significant trends.
Monitoring a bridge like the San Francisco Golden Gate might take 10,000 nodes, says Hartwell. Figure a million or so for a big business application, such as cargo shipping. To enervate the Earth, about a trillion should do the trick. At that rate, sensor nodes must cost next to nothing, yet measure everything.
“The sensor node is a challenging integration problem,” says Hartwell. “You have to put the sensor chip with a radio and a battery and a solar cell in a package that’s inexpensive, yet rugged enough to throw by the side of the road,” he says. As it happens, HP already manufactures hundreds of millions of similarly sophisticated, yet rugged and inexpensive devices today: inkjet printer cartridges.
That experience can be applied to CeNSE nodes: “In both cases you have a complex chip that must be exposed to the environment—to measure it or to squirt ink onto it—and packaged into a integrated unit,” says Hartwell
HP Lab’s memristor technology will also play a key role in getting that integrated sensor node into a tiny package: “The memristor is about doing memory and logic in a technology that’s so small and so low power that a pushpin-sized sensor starts looking like something you can really build,” he says.
Sensor nodes, however, are only part of the challenge of CeNSE.
“How do you capture and use all that data?” asks Hartwell. At a typical data rate, one million sensors running 24 hours a day would require 50 hard disks running in parallel to capture the 20 petabytes of data created in just six months. “The amount of data we’re talking about here is ferocious,” says Williams.
Then it has to be crunched to extract meaningful information. No matter how many gigabytes of data a smart highway might deliver, for example, “you’re only interested in one bit when you walk out that door,” says Hartwell. “Just one bit: Which interstate highway will take you home fastest? If it saves you 20 minutes on your commute, that one bit is worth a lot,” he points out.
HP is approaching sensing networks not just as sensing or moving data or crunching it, but from a holistic perspective, says Hartwell. “We have the networking expertise in our ProCurve division, we have consulting and integration through our Enterprise Services division (formerly EDS),” not to mention business intelligence, storage and data center technologies. Williams agrees: “We’re the only company approaching this from soup to nuts.”
Listening to Earth
CeNSE’s first applications will make living on the planet safer and more convenient. But as the network grows, the breadth and detail of information it gathers could be critical to Earth’s survival, says Hartwell.
“If we’re going to save the planet, we’ve got to monitor it,” says Hartwell. “We have to understand how we’re impacting the planet,” he says, pointing out that we don’t understand how wind farms may affect rainfall or how a cooling sea changes wind currents. Hartwell imagines people volunteering their sensors to feed data to climate change models, just as unused compute cycles are unfolding proteins and unraveling genomes today.
On an individual level, sensing could help people make everyday lifestyle changes: “We have to use this capability to figure out how to change the way we do things: You can tell the kids to turn off the lights, but it’s going to be a lot more effective if the lights turn themselves off.”