For the last few years, a new wireless technology has been waiting in the wings. With promises of Gigabit speeds, it is faster than anything currently in use. Operating by a radical principle that turns today’s radio technologies on their heads, ultra wide band (UWB) is set to hit the big time and products could be out as soon as next year.

UWB – which sends pulsed signals across a wide spectrum – could be so much better than Wi-Fi that some enthusiasts reckon it should replace it in wireless LANs. The majority of the UWB community feel that the weakest target is the cable replacement technology Bluetooth. UWB could beat it hollow, as well as replacing fast cables (like USB and FireWire) that Bluetooth could never touch.

The main factor in deciding where to point UWB is regulation. The technology presents the radio spectrum authorities with systems unlike any they have met before, and their answers will decide what it can do and where.

UWB for beginners
UWB was for a long time known as “impulse” radio; instead of modulating a signal on a particular carrier frequency, it uses short pulses which are made up of a lot of frequencies. These are detected and decoded by pulse receivers, and can carry a massive amount of information. It has also been known as “carrier-free” radio.

A theoretical possibility for some forty years, UWB was picked up by the military as a way to send messages undetectably (it can also allow radar to penetrate walls and deep into the ground). In 1994, it was declassified and, since then, a cluster of companies have been trying to commercialise it, among them Intel, Texas Instruments and Motorola, and a host of smaller start-ups including PulseLink, Xtreme Spectrum, Time Domain, Staccato and others.

The potential is huge: Gigabits per second have been achieved over short distances, while 25Mbit/s over 60 miles has been proposed for military systems using quite low power. These characteristics could make an unbelievably good network for business data.

However, regulators have found it hard to fit the technology into their schemes, which are based on licensing parts of the radio spectrum for separate uses. UWB sends small bursts of energy across a wide spectrum, but it can do so using such low energy that other devices would only see the UWB pulses as “noise”. How can you license a technology that gently touches every wavelength?

In 2002, US regulator the Federal Communications Commission (FCC) came up with a compromise. UWB devices would be allowed, as long as they limited their spectrum to a specific range and kept the radio noise down to less than that emitted from consumer devices like toasters. UWB will have to be quieter than devices that are not radios and it will only be for use indoors.

So the industry has a new toy. Very fast radio transmission that looks indistinguishable from the electromagnetic noise leaked by consumer electronics. But what to do with it?

Wireless that is faster than cables
UWB could leave Wi-Fi in the dust, and make a really fantastic wireless LAN. Even under the FCC rules it could still be significantly better than Wi-Fi. But the competitive issues have warned off many vendors. “UWB looked a decent option, but once 802.11b had taken off, we realised we couldn’t compete with its economies of scale,” says Mark Bowles, vice president of marketing at Staccato Communications.

So, while Wi-Fi aims to make a radio alternative to the Ethernet cable, the majority of the UWB players are focusing on replacing all the other cables, which Ethernet has never been able to replace. The other wireless contender here, Bluetooth, is less well established and cannot replace connections like USB, because it does not have the performance. UWB, on the other hand, has performance and to spare.

A 480Mbit/s short-range (one metre) UWB standard is under development at IEEE’s 802.15.3a working group on wireless personal area networks (WPANs). “That’s three orders of magnitude better than Bluetooth,” says Bowles. At 10m it will still run at 110Mbit/s. It will take a few years for UWB-based 802.15.3a to remove Bluetooth, says Bowles, but it will happen.

While 802.15.3a has a power of 100mW, an even lower power version, 802.15.4a, is proposed, which uses 1mW and has a data rate of 1Mbit/s.

However, the standards effort has not been without problems, as a battle developed between Intel and Texas Instruments’ MBOA proposal and the Xtreme Spectrum solution backed by Motorola. Bowles, firmly part of the Intel camp, reckons this is a classic fillibuster, not a technology debate, and accuses Motorola of packing the meetings: “It might take us till March to overwhelm them, but we will win.”

The next step beyond that is for the USB Implementors Forum, currently looking for a wireless version of UWB, and the equivalent group for FireWire, to adopt 802.15.3a. “The USB group could adopt UWB as a wireless PHY (physical layer)” in 2004,” says Bowles.

But what about LANs?
All cut and dried? Not necessarily. There are USB idealists still fighting for the right to do wireless LANs. PulseLink is the noisiest of them, having just completed a demonstration of 110Mbit/s data transmission, at 20m, and – by going beyond the FCC power limits – at 100m.

However, even at FCC-compliant data rates, PulseLink can do 7Mbit/s at a distance of 100m, which is better than 802.11g, says Bruce Watkins, chief operating officer of PulseLink. Even this hobbled UWB-based LAN would be worthwhile, he reckons.

And PulseLink reckons it has a way to quickly increase speeds if the regulations are relaxed. “We have a software-defined UWB radio,” says Bruce Watkins, chief operating officer of PulseLink. “The same chip can do a 100Mbit/s WLAN or a faster PAN.”

The magic figure with this sort of technology is cost. The technology has to be miniaturised, and a market developed, to hit the magic figure of $10, at which point, it will be put into, more or less, every device. At present, PulseLink’s radio is actually on two chips, making a unit that will sell for $25 by the end of 2004. A single CMOS chip costing $10 should be around in 2005, says Watkins.

Even Watkins doesn’t plan to take on Wi-Fi: “That’s not in my business plan,” he says. “We are taking a longer term approach. We can do things Wi-Fi can’t do.” Among these is the idea of using UWB signals down cables, including electrical wiring and broadband cables.

The future
So, how soon do we get to use it? So far, the US is the only country to have approved UWB, but the UWB vendors all expect Europe could follow, especially if devices are being built which use it and the FCC’s licensing approach can be shown to have worked.

“Europe does things properly,” says Bowles. “While the FCC has politicians, European bodies have technologists.” Although this can result in slower, more methodical work, he expects Europe to follow suit in two or three years. Korea and Japan, he says, are expected to approve UWB soon. As with Bluetooth, if we reach a stage where devices routinely have the technology built into them, a country which rejected it would cause problems for import and export of equipment.

So will we get rid of all the cables on our desks? That’s been promised too many times before, especially when Bluetooth was going through its hype cycle. This time, there’s a big difference: we have technology that is actually capable of doing it.