RSS Feed (xml)
Matt Ettus has the sly smile of someone who sees the invisible. His hands fly over the boards of his Universal Software Radio Peripheral, or USRP, snapping them together with an antenna like Lego bricks. Then he plugs in the naked boards to a USB 2 cable snaking to his Linux laptop.
After few minutes of normal Linux messing around ("Takes forever to boot.... Haven't got the sound driver working yet....") he turns the laptop around to reveal a set of vibrating lines in humps and dips across the screen, like a wildly shaking wireframe mountain range. "Here," he explains, "I'm grabbing FM."
"All of it?" I ask.
"All of it," he says. I'm suddenly glad the soundcard isn't working.
Radio is that bit of the electromagnetic spectrum that sits between brain waves and daylight. It's made of the same stuff that composes light, color, electrical hums, gamma radiation from atom bombs, the microwaves that reheat your pizza.
From our perspective, radio devices behave very differently -- a global positioning system gadget doesn't look like a TV doesn't look like a CB set, even if they are all radios. They are single-purpose machines that use small bits of radio spectrum to do very specific tasks -- about as far from the general-purpose personal computer as you can get. But there's no reason they have to be.
Most of the required components of a radio are the same and can be generalized. And with Moore's law making processors fast enough, much of a radio's function can be done with software.
Building a general radio that can receive and transmit, and attaching it to a software system that can fill in the gaps of what we normally think of as radio, is kind of like the Enterprise's deflector dish: Give engineering 20 minutes and it can do anything the captain needs to move the plot along. One of Ettus' USRPs, with the right daughterboards and radio software, can capture FM, read GPS, decode HDTV, transmit over emergency bands and open garage doors.
The GNU radio project was the brainchild of Eric Blossom, who wanted to create a software HDTV receiver in advance of broadcast flag legislation limiting what hardware was allowed to receive the high-def signal. "We'd just go build one of those things (in software) and moot (broadcasters') control over the hardware," says Blossom.
He teamed up with Ettus, but they lacked a radio platform that was cheap enough to get into many people's hands. They could do a lot with the computer, but there were limits. "How do I get from the antenna into the computer?" explains Blossom. "The computer wants digital samples to work on."
Ettus secured National Science Foundation funding through the University of Utah to design what would become the USRP. "Basically we proposed the 85 percent solution for 10 percent of the price. Given that part of the NSF's charter is about education ... you can get 10 more things in your students' hands for your dollars," says Blossom.
Ettus was drawn more to the technical challenge than the political project.
He wanted to build the HDTV receiver "because it was the Mount Everest ... it was the biggest receive-only mountain." Decoding HDTV was a political act of radio, but, mostly, Ettus wanted to see if he could do it.
Four years later, Ettus hasn't just decoded HDTV, but has gone on to write software that does far more. He's quit his day job to build and sell the USRP hardware full time -- you can buy it from his website starting at $550 for the motherboard.
Ettus and Blossom's software-defined radio on the cheap is popping up in unexpected places, describing a very different radio world from the centralized model that has dominated radio history.
"Decentralized controls enable innovation at the edge -- it's closer to the computer model," says Blossom. "I think what we'll find is that people will come up with things we never really thought about."
Ettus is more concrete about the project's possibilities. Citing Wi-Fi as an example, he envisions "a world in which bandwidth is not an issue. People will create applications that will use that bandwidth, like complete telepresence."
Ettus paints a picture of radio bringing about a many-to-many revolution, like blogging, but for a wider segment of the world. "It enables everybody to be a broadcaster," he says.
Toby Oliver's business is a great example of the street finding its own use for stray radio waves. His company, PathIntelligence, uses the USRP and GNU Radio to track foot traffic in U.K. shopping centers.
Listening for the control-channel signals of mobile phones allows the PathIntelligence setup to pinpoint the location of a phone using triangulation by measuring the difference in time it takes for the signal from a phone to get to multiple antennas.
This works like a very local version of GPS, allowing shopping-center owners to see what shop windows are most popular, and where people tend to congregate or avoid, without actually intercepting any personal data. It's something that processing speed made possible, and the GNU Radio/USRP project made cost-effective.
"Only recently, in the last 12 months, has computing power enabled me to do what I need to in general-purpose software without the expensive development of dedicated DSPs (digital signal processors)," says Oliver. "It means that a whole world of opportunities for tinkerers like me is being opened up."
A person without a phone is invisible to his system, but with the market penetration of mobile phones in Britain, the occasional outlier doesn't damage the data set much. Shopping centers are showing a lot of interest in the information.
But despite his new job, Oliver's background isn't in radio. "In some ways, (software-defined radio) enables the arcane world of RF (radio frequency) to be available to software developers. So you will start to be able to do more and more 'mashups' to RF," he says.
The USRP is being put through its paces in research labs and amateurs' basements all over the world. Ettus sells to companies and governments. Some radio sets out there do more, but Ettus claims that generally the USRP costs a tenth of other software-defined radio-ready equipment. He continues to work on the USRP, developing better signal intelligence and more diverse daughterboards to tune to different bits of radio spectrum.
Blossom is working on a passive radar system that will require a more sensitive hardware setup than the current USRP. His passive radar reads in the ambient radio waves from existing sources, like FM stations and cell towers, and uses them to build a map of the area. At the end of his research, he plans to have "this little gadget that you can plug into a laptop and see what's flying around. We're hoping to see stuff on the order of 50 to 70 kilometers away."
Neither Blossom nor Ettus can predict how their next projects will be used. But that's the point.
Originally posted by Quinn Norton on 06.05.06 at www.wired.com
Read more!
Posts
