X

These walls (and teddy bears) have eyes

From keeping track of babies to spying on people who stick their heads in trash cans, Intel researchers are on the job.

Michael Kanellos Staff Writer, CNET News.com
Michael Kanellos is editor at large at CNET News.com, where he covers hardware, research and development, start-ups and the tech industry overseas.
Michael Kanellos
5 min read
SANTA CLARA, Calif.--Intel Research has developed the ultimate baby monitor for neurotic parents.

The experimental system--which consists of a series of sensors under the baby's mattress and a camera mounted on a wall--will monitor a child's heart rate, temperature and movement; stream video of the infant; and even take pictures. Captured data is sent to a parent's PC.

Though the system is geared mostly toward providing parents with better information about their kids, there is an entertainment aspect to the monitor as well. The pictures taken nightly by the camera can be strung together to form a composite video charting the baby's development. A prototype teddy bear developed by the same group, meanwhile, contains a hidden video camera.

"Clearly, there are privacy issues."
--Brooke Foucault, anthropologist

"People are telling us they want more spontaneous images," explained Brooke Foucault, an anthropologist who tested the monitoring system on 25 families. "Clearly, there are privacy issues."

Remote observation was one of the dominant themes at the company's annual research day on Wednesday at its headquarters here, where it unveils a number of conceptual uses for high technology. For the past several years, Intel scientists have theorized that the next wave of computing will involve designing "proactive" machines that can harvest and manage information for overwhelmed human beings.

Whether one sees these systems as a useful exploitation of technology or an insidious step toward Big Brotherdom is in some respects a matter of implementation, personality and opinion. In any event, the privacy issues are being examined in detail.

In another experiment, researchers have tagged all of the items in a person's house with RFID sensors that effectively will tell a remote computer whether the occupant has moved a spoon or turned on the television.

Though it might sound Orwellian in the abstract, the system is being designed to provide relatives or professional caregivers information on the daily habits of the elderly.

Though it might sound Orwellian in the abstract, the system is being designed to provide relatives or professional caregivers information on the daily habits of the elderly. Did they take their medicine? The system can't precisely confirm that, but it will relay information that a specific pill bottle was moved at a particular time.

"We assume they took their medicine because these people are doing this so they won't have to go into assisted living," said Sunny Consolvo, an Intel researcher. Nonetheless, privacy is taken into consideration. Test patients say they prefer the RFID reader to be embedded in a bracelet rather than a wall because they can take the bracelet off to halt the tracking process.

Another experimental technology, called Robust Video Superresolution, seeks to improve the quality

of video and still images captured on small digital cameras and cell phones. RVS examines successive frames from a low-resolution camera, studies how the different objects in the successive frames move, and then creates a composite image of what the original tape might have looked like had it been taken with a high-resolution camera. The system relies on Bayesian probability to create the extrapolated, higher-resolution video. This will make those surreptitious health club videos more crisp, but also the vacation video.

"We squeeze out the remaining information," said Horst Haussecker, an Intel researcher. The system, which may appear in consumer cameras or PCs in three to five years, actually derives from semiconductor manufacturing equipment. In chip factories, these techniques are used to determine the existence of defects in transistors or other structures that are too small to examine directly.

"The idea is that the images start to tell a story about the kind of things that matter to people."
--Eric Paulos, senior researcher, on images from a camera installed in a garbage can

Another experiment, meanwhile, shows how Wi-Fi can be used to track individuals inside buildings, where Global Positioning Systems won't work. In this system, a Wi-Fi base station times how long it takes a signal to reach a notebook and for the notebook to send back a return signal. By charting the time that transaction takes, the base station can determine where a person is. It is accurate to about a meter, said Senior Staff Architect Stuart Golden.

"We're starting to talk to standards bodies" about adding the technology to a future Wi-Fi specification. The person holding the notebook, he added, can also at his or her option cut off the tracking signal and still get Internet access on the notebook.

On the lighter side, Eric Paulos, a senior researcher, touted his Jetsam project, a garbage can equipped with a digital camera and a projector. When someone throws garbage into it, or just sticks their hands past the camera lens inside the receptacle, the camera takes a picture and beams the image, along with earlier images, onto the sidewalk.

"We had a lot of sticking the head in," he said during an experimental installation in Berkeley, Calif., last year. "The idea is that the images start to tell a story about the kind of things that matter to people."

Paulos, who has been a computer researcher for most of his career, also had his team spread 1,200 matchbooks around sidewalks in Portland, Ore., asking individuals to send in SMS messages to vote on whether they loved, or hated, the city. In another stunt to seek out human behavior, his group scattered prestamped postcards (with fictional messages to a fictitious recipient) around San Francisco. About 10 percent to 20 percent came back.

Researchers present at the event all largely cautioned that there is no guarantee that any of the experimental technologies discussed will ever come to market. Instead, the idea is that concepts or usage patterns discovered during experimentation will lead to new applications for computing. Some groups, for instance, have expressed interest in the baby monitor to study the sleep patterns of chronically stressed adults.

Engineers will have to foolproof their systems as well.

"You also don't want to reboot," Foucault said of her baby monitor. "The last thing sleep-deprived parents want to do is fiddle with sensors in the middle of the night."