Next: 'convergence of robot, machine-perception, information technologies.'

EE Times: When did you first become interested in wireless?
Joseph Mitola: I guess it started when I was 11 years old, when I built my first radio with one of those crystal-radio kits with a hearing-aid-type thing for your ear. In high school, my father gave me a Datamation book on how to build a flip-flop, and I decided that would be pretty interesting to build.

EET: You're called the father of software-defined radio, or SDR.
Joseph Mitola: Well, I wrote the first technical paper on it in 1991, and I had already developed one for the Air Force. GEC/Marconi had also done some work on it. In 1995 I became editor of IEEE Communications Magazine and put together a series of articles on wideband analog-to-digital converters and digital signal processing. I wrote the lead article on architectures. That was a watershed event, as it brought together research from around the world. After that paper appeared, I got invited to be chair of the SDR Forum.

EET: What's your definition of SDR?
Mitola: It's about the coverage of the radio. Two-to-two [2 MHz to 2 GHz] was the original definition, but does a handset do all that? Not really; it's more partial-band coverage.

EET: Is SDR too expensive?
Mitola: That's self-interest talking. What motive do I [speaking as a handset maker] have for selling multimode handsets? I can make a lot more money if [users] keep upgrading one band or radio at a time.

EET: But isn't power an issue for SDR?
Mitola: You don't need it for single-function devices like a pager, but you do need it for emergency management. You need bridges be­tween agencies. Also, for high-volume markets, you have to optimize, since one-zillionth of a watt [extra] can put you at a competitive disadvantage. It's a matter of which technology in which market at what time. It's all business-driven, not technology-driven.

EET: Why is SDR viable now, when it was once considered an exotic technology?
Mitola: It's a combination of improvements in analog-to-digital converters on the re­ceive side and wideband linear modulators on the transmit side, all powered by Moore's Law.

EET: What about the analog front end?
Mitola: This is where the physics is really daunting. The work that has to take place involves digitally controllable analog circuitry—using MEMS [mi­cro­electromechanical systems]. For example, you could have a capacitor based on interdigitated fingers, and these would be moved by a piezoelectric motor measuring 10 x 60 microns to change the [value of the] capacitance to change a tank circuit. MEMS had issues such as cycling, thermal management, yield, etc., but we're moving beyond that now. How­ever, true portability of waveforms is still a vision with challenges that need to be overcome.

EET: Define cognitive radio.
Mitola: I describe it as a really smart radio that would be self-aware, RF-aware, user-aware, and that would include language technology and machine vision along with a lot of high-fidelity knowledge of the radio en­vironment.

When I coined the term cognitive radio, I said at least 20 years [until takeoff]. Little did I suspect that the FCC would adopt the technology before I was even done with my dissertation. I would say that one of the things to think about [next] is the convergence of robot technology, machine-perception technology and other information technologies. I think that's a bit of a sleeper.