SAN FRANCISCO A blue-ribbon panel convened here at the International Solid State Circuits Conference to ponder the rather Shakespearean question, "To UWB, or Not to Be?" The answers were decidedly mixed.
Panel chair Thomas Lee, a Stanford University professor, opened the discussion with a loose definition of UWB radio: a radio that, when measured as a system-including the antenna-had a bandwidth of at least one quarter of its center frequency, or, operated above 6 GHz with a bandwidth of at least 1.6 GHz.
Lee also pointed out that the biggest impediment to UWB was its environment. In the GHz region RF energy is quickly absorbed by air, with particularly prominent absorption peaks for water vapor, oxygen and most solids. "It's best to avoid locations where there are atmosphere, rain or trees," Lee suggested helpfully.
Don Cox, and fellow professor of Lee's at Stanford, quantified Lee's concern. "I hear people start out taking about UWB for measuring distances, or for very local communications," Cox said. "But in the next sentence they are talking about hundreds of meters and Gigabits per second. Sorry, but that is not in the cards."
Cox showed an established relationship between possible data rate and distance that showed a 1/d-squared relationship in open air, and a 1/d to the 3.6 power relationship in a building environment with interior walls and doors.
He pointed out that increasing power or adding antennas could change the scale of the y-axis the data rate but not the relationship. UWB was still limited to very short distances and moderate data rates.
Bob Broderson, a University of California Berkeley professor, added his own note of skepticism. He pointed out that while the FCC had made a great deal of spectrum available, there was no reason to believe that current frequency-hopping approaches to exploit it could provide high data rates at a reasonable cost.
Broderson suggested that rather than continue to struggle with very high-performance OFDM wide-band architectures, it might be better to explore pulse radio ideas more fully. "This might result in an approach no one has thought of yet, using very low transmit power and lots of processing gain at short range," he said.
Before the mood turned too dark, Jeff Foerster, a researcher at Intel's Corporate Technology Group, sounded a note of optimism. While admitting that neither UWB nor anything else was a universal solution to the problem of ubiquitous connectivity, UWB might well be part of the answer, even if only at ranges below 10 meters. He suggested that clever design might altogether eliminate power amplifiers, and could significantly simplify receivers.
Before the optimism could get out of hand, however, Ali Jajimiri, a professor at the California Institute of Technology, jumped on the notion of pulse radio with both feet. He pointed out that pulse transmission inherently required high voltages, a very broadband antenna and a highly linear receiver that could tolerate a wide noise bandwidth. "We pretty well have to give up omnidirectional antennas in favor of multiple-antenna systems to get sufficient antenna efficiency," he suggested.
Next, Keiichi Ohata, principal researcher at NEC, lambasted UWB without a word of criticism. Instead, Ohata outlined development work at NEC on a 60 GHz narrow-band data link using ASK modulation to achieve 1 Gbit/s data rates at 100 meters. He said that because the link was effectively blocked by walls and was inherently short-range, it was highly suited to indoor, private picocell arrays. Ohata said the work was in line with the 802.15 mm WIG standards effort.
Finally, Ian Oppermann, director of the Center for Wireless Communications and professor at Oulu, Finland, took on one of the criticisms of UWB. "The European Union is making a major investment in what they perceive is the need to catch up with the US in UWB," Oppermann said. "But one of the big concerns is interference. The EU is very conservative. Interference with the operation of existing systems would kill UWB in Europe."
Oppermann described a series of experiments in which his team had constructed a group of extremely powerful UWB transmitters through a misunderstanding, they turned out to be 30 dBM more powerful than intended, he admitted and aimed them at conventional wireless devices to measure interference. Surprisingly, he found that the equivalent of thousands of UWB transmitters operating only centimeters from an 802.11b transceiver caused only a minor degradation in data rate, and that by the time the sources were a meter or so away their effect was negligible. Bluetooth devices, he found, were essentially unaffected at any range.
In sum, the panel seemed to agree that UWB was feasible, though it might take revolutionary developments in pulse radio to make it so. But the august group did not seem to find any particular use for which it might be good.