In today’s culture, wireless rules. WiFi-enabled laptops easily connect to Internet hotspots, prevalent in most cities, and cell phones are virtually ubiquitous. But while much of the technology in these devices is advanced, the radio chips that transmit and receive those wireless signals are, for the most part, still based on an older, highly inflexible model: they’re hardwired with access to only a single kind of wireless frequency, such as WiFi or cellular or GPS.
In recent years, researchers have been working on a wireless chip that can do it all: tune into television and radio stations, GPS, cellular signals, WiFi, and WiMax–or even unlock your car door. The trick is to use software to modify the radio hardware, a concept known as software-defined radio (SDR).
Some of the pioneering research on SDR is happening at Lucent’s Bell Laboratories, a facility that’s invented a number of the wireless technologies used today. Bell Labs’ researchers are now developing SDR for base stations, the wireless hubs that communicate with devices.
Currently, it’s expensive to deploy and install the base stations that route wireless signals from, for instance, Cingular, Verizon, or Sprint, to their customers. And when the station’s wireless technology needs to be updated, the companies have to redeploy and reinstall the radio technology. By using software-defined radio, however, the hardware could be easily and inexpensively brought up to speed by simply tweaking the software. In other words, these more-flexible base stations could be upgraded quickly–possibly leading to fewer dropped calls, more reliable signals, and extended ranges.
Tod Sizer, director of broadband wireless research at Bell Labs, and a speaker at Technology Review’s Emerging Technology Conference today, discussed software-defined radio with our editors.
Technology Review: The term “software-defined radio” is kind of clunky, and not too descriptive to people who aren’t familiar with it. What’s your cocktail party definition of the technology?
Tod Sizer: It’s the ability to take a radio [in a cellular base station] that is doing an application that you define today, and tomorrow you can reconfigure it to do something different. That’s one model of software-defined radio. There’s another model, and it’s that you have a single radio [in a handheld device] that has both Bluetooth and [Wi-Fi] 802.11, for instance. We know what those standards are today, but you want to have one radio that can do both. That’s about reconfiguring immediately where the radio can either be one or the other. A lot of software-defined radio has also been pushed strongly by the military, which wants a radio that can communicate with different types of radios, such as those used by the Marines, the Coast Guard, or firefighters. It’s a very important problem and one that the military has pushed as they’ve been a supporter of software-defined radio.
TR: Why is software-defined radio exciting to you?
TS: I think it’s exciting for a number of reasons. One is that it allows pure researchers a tool to try very different things very quickly. Say we have developed a slightly different coding technique. In the past, we’d have to wait for it to be coded [in hardware], and then we would have to wait a year and a half to see if our implementation was correct. Today, if we have an idea, we upload it into our vans and test it immediately. It allows us to test ideas very quickly. To be able to very quickly go to the field is important.
And from a commercial perspective, it’s good for my customers, including Verizon, Sprint, and Cingular. We’d like to help them improve systems that they’ve already deployed over time.
TR: Where is the technology right now?
TS: I think it’s still in its early stages…Some of the programmable platforms aren’t quite where they need to be in terms of cost performance. I think the opportunities are there today, and if there were devices were cost competitive, it would be quite prevalent. We’re still searching for a device that’s cost effective and has the performance. Some of these devices used for the software-defined part of the radio are quite expensive. There’s the rub.
TR: It seems like software-defined radio can fall into two categories: the base station and handheld devices. What are the economic challenges of both?
TS: A hardware installation is reasonably expensive and you want to have it last a long time. So having it reprogrammable and reconfigurable is important for my customers. Having flexibility is important and is very well suited for software-defined radio. You can justify the initial expense because you’re going to be providing expanded usage for a longer period of time, whereas with handsets it’s very different because their cost is cheap. Adding cost to it so that three years down the road you can add something to it that you didn’t know about isn’t cost effective. I think there are places and applications, such as base stations, where price points are higher, where we’ll see software-defined radio early. And in handsets, where cost is king, we’ll likely see it much later.
TR: How is software-defined radio technology making its way into commercial applications?
TS: Certainly we’re seeing reprogrammability in base stations today. Vanu is a company that has a base station that’s programmable today. There are products with different levels of software-defined radio.
TR: Once software-defined radio becomes more prevalent, how do you see it changing the landscape of wireless communications?
TS: There are various applications where software-defined radio could play a larger role; one area is cognitive radio. [Cognitive radio is an offshoot of software-defined radio that lets devices scan for wireless signals, from broadcast television to WiFi, and find and employ unused parts of the spectrum to allow more efficient use of the radio spectrum (see “Cognitive Radio,” March/April 2006) –TR.] There’s some talk about taking advantage of the unused television spectrum by radios that know about the environment that they’re in. That cognition is something that we’re looking at quite closely. A model like that is the only way to use that spectrum efficiently–and the only way to be flexible enough to do that is with software-defined radio.