Feature Story

THE QUEST FOR UNIVERSAL RADIO

Where are they now? Many new stars get their 15 minutes of media attention, then disappear - and are found years later by nosy reporters on a slow news day, pumping gas in Phoenix. Some technology concepts also achieve media darling status, then fade into obscurity. For example, what fate befell software-defined radio (SDR)?

The concept entered the techno-babble lexicon a few years ago, accompanied with buzz about how it would free the masses from their bondage to their carriers' hardwired cellular handsets and networks.

"The vision was to create a universal translator for radios," says Carmi Levy, senior analyst at the London-based Info-Tech Research Group. Instead of custom hardware with dumb chips that can only handle a single signal - Wi-Fi, GPS, Bluetooth or GSM - one smart chip connected to an antenna would do it all in the same device. Consumers could have a PDA, BlackBerry, cellphone, Internet device and remote control for the house all in one handset. Wireless providers could save vast sums of money by downloading software patches to upgrade to next-generation offerings instead of physically upgrading millions of handsets.

This SDR dream system, as originally conceived, has proved more elusive than imagined. It's technology that's easier said than done, says Levy. But the successful components of SDR have quietly infiltrated cellular infrastructure, and other elements are evolving in surprising new ways.

Where is SDR today?

Most wireless players are involved in SDR research and development, but none have disclosed the status of their efforts, says Levy. "I would say all the major carriers, handheld vendors and wireless players are involved to a certain extent - but good luck getting them to confirm they're actually using SDR."

In the ferociously competitive wireless sector, providers are jealously guarding their SDR secrets. "I believe most cellular base stations already use SDR today, as it allows wireless providers to handle different handsets," says Al Margulies, executive director of the SDR Forum based in Denver CO. "But intellectual property issues are so strong that I can't say who's involved."

The idea of SDR has its roots in the US military, which has been doing pioneering research since the early-'90s. In 2001, the Joint Tactical Radio System (JTRS) initiative was started up to work on the problem with other participating nations such as Canada, Japan, Sweden and the UK, says Levy. "The goal is to develop one radio standard that cuts across national and service-based boundaries," he says. "This is largely the result of the US' experience in Iraq and Afghanistan, where the army's radios don't talk to the navy, air force and so on." JTRS is testing prototypes, but it'll be about five years before we see anything workable, he adds.

Levy notes the military has much deeper pockets and longer horizons than the commercial sector. "When the military gets involved and puts billions into evolving a technology, it's a given that if the project is successful, it'll have massive implications for the commercial sector," he says, pointing to the evolution of GPS as a precedent. It too was once a military program, but once it was in place for military applications, the commercial sector piggybacked on top of that early success. "Now GPS is ubiquitous, and you can buy a receiver for 200 bucks."

But there are design constraints in developing SDR handsets for the mass market that don't exist in military applications. "A soldier's radio can be large, consume a lot of power and be expensive - the military can live with these factors," says Louis Bélanger, CTO of Lyrtech Inc., an advanced digital signal processing (DSP) provider based in Québec City.

Not so in the commercial sector, where efforts to create consumer handsets have been stymied by a number of conflicting design factors. "The issue in handsets is the sensitivity of size, weight, power and cost. The fact that it's programmable technology means it costs more and takes more space than handsets in use now - it's difficult to cram it all in," says Margulies, adding that these constraints don't exist in SDR-based base stations.

There are also the laws of physics to contend with in the development of antennae that can handle a range of signals, he says. "Devices have to be tuned. If you look at a satellite antenna, it's a big dish; on a cellphone, it's a stub; and on a car, it's a whip. These have different operating characteristics at different frequencies, so the power output chips and getting them to operate across bands is difficult."

Nevertheless, Margulies doesn't believe the current inelegant methods of mass-production and dissemination of physically upgraded devices to consumers are so firmly entrenched they can't be dislodged. "No, SDR is inevitable," he says, pointing out there are powerful incentives in the wireless sector to crack the problem: the ability to get new designs to market fast and cheap, and to maintain devices in the field remotely.

SDR morphings

But consumer demands aren't standing still - these too are evolving and influencing the direction of SDR technology. "The original dream was to have a human radio conversation where if I'm talking to someone, I could switch and talk to someone else using a different protocol but modulated on the same radio," says Bill St. Arnaud, senior director of advanced networks at Ottawa-based CANARIE.

This single-channel approach is already seen as limited, he says. "Then people discovered, no, they wanted to carry on five simultaneous conversations - while talking to someone, you could also be getting GPS co-ordinates, communicating with your computer and so on. The idea of a multi-receiver radio has really caught on in the past five years."

This is the idea behind a new buzzword - cognitive radio - where multiple channels are used to allow the radio to multi-task. "The receiver needs to be in constant communication with multiple sources, not one," says St. Arnaud, adding this will allow the radio to sense and adapt to whatever environment it happens to be in, or hunt for signals based on whatever criteria the consumer chooses - the best Wi-Fi signal, the cheapest wireless phone rate and so on.

While the true SDR vision of a smart chip that can process different signals hasn't been fully realized, many of the techniques used in SDR research are being used to tackle this new challenge, he says. Researchers are still constrained to use custom chips for each type of signal in their cognitive radio prototyping. "They're taking the successful parts of SDR to make simpler chips. Devices have five chips but they now off-load traditional analogue hardware processing to a central SDR processor. It really simplifies the job by having a radio frequency (RF) front-end with a specific chip for each channel and at the back-end, a common chip, the CPU in the radio, that does all the signal processing," he says. "So it's a hybrid - they picked the best ideas from SDR and combined it with multi-channel concepts, and this is the way it's evolving."

While research on pure SDR is ongoing, there are other contending technologies to solve the interoperability problems of military and public safety applications, St. Arnaud says. "But I wouldn't call them competing technologies as they use SDR tools and techniques." For example, an alternative to spending billions developing a true SDR radio could be a multi-channel cellphone. "You could instead give people cellphones with Wi-Fi and Bluetooth - something off the shelf that has eight or nine services they can access. It's more effective and cheaper than a multi-billion dollar boondoggle."

St. Arnaud doesn't believe true SDR is likely to materialize, but hybrid, multi-receiver SDR is already happening. "I think we'll see a wave of new multi-receiver cellphones in the near future - but consumers won't be aware they're SDR-based, just that they have neat features."

Meshing is one such feature, where a radio can act as both tower and antennae using Wi-Fi. "When you're not using your cellphone, it could be acting as a router, redirecting traffic within 20 feet to other phones that don't have good reception. This would be a big boost to carriers in elevators or other poor coverage areas."

Liquid radio is another new SDR-based concept and buzzword, he says. "This can do processing as close as possible to the RF stage, which allows incredible flexibility in bandwidth. If you're in the right area, you could download HDTV through your cellphone to a computer, so the radio becomes a sophisticated device not limited by narrow voice band."

While there's no consumer clamour for SDR, there is plenty of demand for features such as liquid radio that are enabled by it. "Right now, the big problem on 3G cellphones is the high cost of data," St. Arnaud says, pointing out that downloading a movie to a cellphone is prohibitively expensive. "With liquid radio, you could pick up signals from various sources on multi-receiver radios, then have the radio feed the movie to a computer or DVD in the car for the kids. You could be getting a high volume of data at low cost and not be locked into a single provider."

More bandwidth is key in the evolution of cognitive radio and its offshoots, says Lyrtech's Bélanger. "The way they want to use cognitive radio in the commercial space, it gives you more bandwidth," he says, noting there are trials underway in Japan and Europe to free up unused RF spectrum for high bandwidth communications. But SDR-based devices won't be able to use free spectrum unless the unused chunks are identified at their end to enable the interaction. "You can't just lock onto free spectrum - you need an RF resource manager to define, organize and allocate RF." Yet another SDR-based hybrid concept is born for this area. "So it'll be semi-cognitive radio."

Return to: 2007 Feature Stories