[MLB-WIRELESS] UWB: Bounced message from Marian Szczepkowski

Wed Jul 3 09:51:57 EST 2002

People may be interested in this.... it was originally posted by Marian
Szczepkowski, but was bounced due to it's size...

        01 Jun 02  USA: ULTRA-WIDEBAND WIRELESS - FAT PIPES FROM THIN
        AIR? - UWB DOESN'T BREAK THE LAWS OF PHYSICS, BUT IT'S AN ...
         (NETWKM)

By Andy Dornan.

Ultra-Wideband Wireless - Fat Pipes from Thin Air? - UWB doesn't break
the laws of physics, but it's an exciting technology-and it could even
live up to Bluetooth's hype.

Up until now, you couldn't take the phrase "mobile Internet" literally.
Whether describing the slimmed-down, text-only services available
through a smart phone or full-scale wireless Web surfing, mobile
Internet has really meant mobile access to the Internet.

If current tests of Ultra-Wideband (UWB) technology are successful, we
could soon see the real thing: tiny routers embedded in every electronic
device, and eventually other objects, from pens to business cards.
Before that comes to pass, though, we'll probably see high data rate
wireless LANs (WLANs), several times faster than current IEEE 802.11
networks. The target speed, already demonstrated in the labs of vendors
such as Intel and XtremeSpectrum (www.xtremespectrum.com), is around
100Mbits/sec-real throughput, and not shared between multiple users.
Perhaps more importantly, these high data rate WLANs could have battery
lives measured in years, rather than hours.

Of course, this comes at a price. Despite what some proponents of UWB
might claim, the amount of data that can be sent over the airwaves is
still finite. Unrestricted UWB transmissions could interfere with
current radio spectrum users, cutting into the bandwidth of existing
wireless, and even fiber, networks. Because UWB is also touted for other
applications, from personal radar to surveillance systems, its
widespread use could even mean a loss of available capacity for voice
and data networking.

Because of such fears, regulators have been reluctant to allow UWB. So
far, the FCC has permitted only four companies to manufacture and sell
UWB devices, and none of the devices are intended for communications.
Three make imaging systems, aimed at construction and safety personnel.
The fourth product, and the only one sold to the public, is a "toilet
ventilation device."

That changes this month, as new regulations announced in February 2002
come into force. Peer-to-peer UWB is limited to low power and is still
only allowed on an experimental basis, but the industry has high hopes.
The tight restrictions could ultimately result in better systems that
use airwaves more efficiently, and with such low energy requirements
they won't need batteries or power cords.

SURFING WITHOUT WAVES

The most extreme hype around UWB claims that it uses no radio spectrum,
and that it eliminates radio's traditional reliance on waves in favor of
short pulses. The first claim is simply wrong: UWB uses more spectrum
than other radio systems, hence the name. The second claim is only
partially true. All networking technologies, wireless or otherwise, are
based on electromagnetic waves, and UWB is no exception. It does,
however, abandon the concept of a carrier wave, meaning a specific
frequency to which radios must be tuned.

Instead, UWB broadcasts on many frequencies simultaneously, distributing
its signal across a vast bandwidth. This is similar to spread spectrum,
the basis of 802.11b WLANs and Code Division Multiple Access (CDMA) cell
phones use, but UWB spreads a signal over a wider band-potentially
across the entire radio spectrum, though no government would permit
this. The FCC, the only regulator that allows UWB communications systems
at all, has initially restricted them to a bandwidth of 7.5GHz (see
table on page 68). That's hardly infinite, but still many times broader
than the spectrum allocated to every other networking system combined.

The huge bandwidth requirement means UWB can't be given a dedicated part
of the spectrum in the same way as most wireless technologies. (Even
unlicensed systems, such as WLANs, are still confined within a
relatively narrow frequency range.) UWB must instead share the airwaves
already allocated to other systems. This reuse of existing bandwidth
leads to claims that UWB provides something for nothing. The theory is
that because the signal is so thinly spread over so many frequencies,
the amount of interference it causes at any one frequency should be
negligible.

Naturally, existing radio spectrum users don't see things that way. When
the FCC first announced that it was considering allowing UWB
transmissions, it received thousands of pages of comments, mostly from
companies, user groups, and even other federal agencies worried about
interference. Broadcasters, airlines, radio hams, astronomers, and the
military are all concerned that UWB's gain could be their loss.

UWB proponents say low power interference already occurs at many
frequencies, permitted by the FCC's Part 15 rules. These rules are
intended to allow devices such as computer processors, which
inadvertently act as transmitters at their clock frequency-currently
anywhere from about 300MHz to 2GHz, which encompasses most of the bands
used by broadcasting and cellular services.

UWB transmitters would stay within the limits set by the Part 15 rules,
but actually transmit useful information. There is already a precedent
for this, in the form of amateur broadcasting: Anyone in the United
States can set up an unlicensed radio station, provided it keeps within
the Part 15 rules. The rules are quite strict, limiting the effective
range to about 100 meters, or 300 feet-impractical for most types of
broadcasting, but not for a WLAN.

This argument doesn't convince existing spectrum users. Most existing
Part 15 devices transmit in only a small part of the spectrum, whereas
UWB needs to cover a wider bandwidth. Although UWB's power at any
particular frequency might be within limits, its overall output could be
higher than that of a Pentium chip or even a Part 15 radio station.
Users are also concerned about the sheer number of UWB devices that
might eventually be produced, each of which could add a little more
interference.

UWB's strongest critics are in the cell phone industry. Sprint PCS
(www.sprint pcs.com) was the most aggressive, because its network is
based entirely on CDMA, the cellular technology most susceptible to
interference. On the basis of tests conducted jointly with UWB company
Time Domain (www.timedomain.com), Sprint claimed that widespread use of
UWB would reduce its network capacity by up to 1,000 people at busy
periods in a typical city-even if each UWB device were restricted to an
output power only 6 percent of the Part 15 limit. It also threatened to
sue the FCC for breach of contract.

Other cell phone companies also attacked UWB, because they all hope to
upgrade their networks to some form of CDMA. Cingular even claimed that
UWB could interfere with pacemakers and hospital equipment. In a January
2002 letter to the FCC, the United States's four largest mobile
operators-AT&T Wireless, Cingular, Sprint, and Verizon-and the CDMA
patent-holder Qualcomm jointly warned that, "People in an office
building trying to use their cell phones to report a fire or other
emergency could well have their calls blocked if there are UWB devices
operating on a nearby local area network."

ULTRA-WIDE BANNED

It's easy to be suspicious of cell phone companies' motives. They could
be attempting to use concerns about interference to shut down a
competing technology, rather than simply ensure that their own
technology continues to work. Sprint wouldn't comment on this, but its
filings with the FCC have claimed that the UWB companies "want to use
Sprint PCS's spectrum for free to provide telecommunications services in
competition with Sprint PCS's services."

The distinction might seem arcane, but it's important legally. Cellular
operators have paid billions for licenses entitling them to use a
certain part of the radio spectrum, so they have a legitimate reason to
be upset if others are causing interference in "their" bandwidth. These
licenses do not entitle them to a monopoly over wireless voice and data
services, however. In North America, and increasingly in other countries
(See "Europe Warms to Hotspots," page 26), anyone can offer a service in
the freely-available spectrum used by WLANs. Though this might upset
cellular operators, an attempt to shut down these operators would simply
be anti-competitive.

The FCC eventually dismissed most of the cellular operators' concerns
about interference, at least with the voice signal itself. It said that
Sprint's tests were conducted in ideal lab conditions that didn't
correspond to the real world and pointed out that people experiencing
interference could simply move a UWB device away from a cell phone, or
switch it off entirely. The cell phone companies disagree, saying that
customers expect their cell phones to work everywhere and are likely to
blame the service provider, rather than their own UWB devices, when
calls are dropped.

Even if UWB doesn't interfere with the voice or data signals, cellular
operators have another legitimate reason to be worried: the Global
Positioning System (GPS). The GPS also uses CDMA, and because its
signals are transmitted from satellites, they can be extremely weak by
the time they reach users. Potential interference with the GPS signal is
also what upset the military and the airline industry, which now depends
on the GPS to route almost all flights within North America.

Cell phone networks rely on the GPS signal in two ways. First, the
narrowband CDMA systems used in the Americas and some parts of Asia
require the GPS to synchronize the clocks at their base stations. The
networks would fail completely if the GPS signal was jammed, but this is
unlikely. Most base stations are in areas with a clear view of the sky,
making the signal easier to pick up. They're also fenced off to protect
people from their radiation, simultaneously protecting the station from
UWB or other potentially disrupting Part 15 devices.

More worrying, the FCC's Enhanced 911 (E-911) mandate requires all U.S.
carriers to track their users' location, and the most accurate way to do
this is to build a GPS receiver into every phone. Because cell phones
have to work indoors and get a fix quickly, their receivers are often
more sensitive to the GPS signal than the standalone GPS devices used by
hikers, aircraft, and even the military.

This extra sensitivity (achieved through updates sent over the cellular
network) also makes the cell phone's GPS receivers more vulnerable to
interference, though exactly how vulnerable has yet to be determined.
Qualcomm (www.qualcomm. com), the largest vendor of these enhanced GPS
systems, has warned that its technology simply won't meet the level of
accuracy required by E-911, if a wideband Part 15 device is nearby.
However, UWB vendor XtremeSpectrum claims that it can build UWB and GPS
into the same device without interference problems.

The FCC is worried about the effect of UWB on GPS, so has heavily
restricted UWB communications within the GPS band-to a level about 0.04
percent of that allowed for other Part 15 devices, which it says should
be below the regular background noise level. (See Figure 1.) It's also
restricted UWB communications within other bands, including those used
by cell phones, though not by as much, and by amounts that vary based on
the type of UWB device.

Within a building, UWB can transmit at up to the same level as the
Sprint tests (6 percent of the Part 15 limit). Outside, they're limited
to one-tenth of that, because walls and other obstacles inside help to
dampen the signal. Higher power levels are allowed for some other
purposes, including imaging, but these levels are already attracting
criticism as too low.

Although UWB is often touted as revolutionary, this only applies to its
applications in communications. As far as imaging is concerned, a
similar system called ground-penetrating radar (GPR) has been used for
over 40 years by miners, archaeologists, Apollo astronauts, and anyone
who needs to see underground. Most recently, it's been used in laying
and repairing fiber cables, which, unlike their copper predecessors, are
invisible to metal detectors. Despite relatively high energy levels, it
hasn't caused any interference problems for wireless networks, because
most of this energy is directed into the ground.

GPR now falls under the same rules as newer UWB systems, causing an
uproar among its existing users. Since the regulations were first
announced, almost every comment sent to the FCC has been from a GPR
service provider or user, warning of dire consequences if it's banned.
Among other things, they say that a GPR ban could lead to an increase in
the "backhoe problem," when someone digging into the ground accidentally
cuts through a piece of fiber and causes a network outage.

FREE BANDWIDTH, FREE ENERGY?

Because of the FCC's different rules for indoor and outdoor UWB
networks, two different types of devices are likely to emerge. The
regulations say that the more powerful devices, intended for indoor use
only, will need a mechanism preventing them being taken outside. The
most obvious such mechanism is a power cord (with no battery), which
seems to negate the main advantage of a wireless system, but might not
make it entirely pointless. Many consumer devices do spend most of their
lives plugged into wall: For example, UWB could stream video from a VCR
to several TVs around a home.

Devices designed for outdoor use are more interesting. Many of these
will probably be used indoors too, but need to meet the stricter outdoor
interference requirements, because they'll work in "unthethered" mode.
Their low power means they're likely to have a short range, more like a
much faster version of the Bluetooth Personal Area Networking (PAN)
technology than a full-scale WLAN. Indeed, the IEEE's 802.15 working
group, which studies wireless PANs, is considering UWB as the Physical
layer for a future standard. The current standard, 802.15.1, is almost
identical to Bluetooth, but the IEEE wants future versions to offer a
higher data rate or lower energy consumption.

UWB can achieve both of these goals, thanks to the way it generates a
signal. Instead of using complicated modulation schemes and antennas, it
emits staccato pulses of white noise directly from a chip. The gaps
between pulses are much longer than the pulses themselves, so for most
of the time the system is idle. Coupled with the FCC's restrictions on
output power, this could enable a device to run for months between
battery charges.

There's even research into eliminating the battery altogether, instead
relying on "energy scavenging," which extracts tiny amounts of energy
from the environment as needed. For example, a wearable UWB transmitter
might be able to harness the energy generated when a person moves, as
some "self-winding" watches already do.

There isn't much energy available for scavenging, so a batteryless
wireless system might need to reduce its power even further. The most
efficient way to do this could be to send a transmission via several
short routing hops, rather than a single longer one. This is because the
received signal strength in an indoor environment falls off rapidly as a
transmitter and receiver move further apart, often by a factor
proportional to the third or fourth power of distance. Even what appears
to be a roundabout route can use less energy than the direct path.

This is the wireless version of peer-to-peer distributed computing.
Whereas the peer-to-peer systems currently deployed across the Internet
distribute storage (Napster) or processing (Web services), this would
distribute routing. Every node would act as a router, automatically
passing packets along the most energy-efficient path. (See Figure 2 on
page 69.) This vision does have some weaknesses: The more links in a
chain, the more likely it is to be broken by devices moving too far
apart. And the most efficient path is one that reduces the distance
between nodes to zero, otherwise known as a wire.

Energy scavenging networks probably won't have high data rates, but more
conventionally powered UWB transmitters promise 100Mbits/sec or more. A
UWB network's data rate is proportional to the number of pulses per
second, so in theory it can scale with the chip speed. Some vendors even
describe UWB as "Moore's Law Radio," saying its performance could double
every year or two.

Such claims are probably over-optimistic, and UWB will still run up
against limits. (See "How Vendors Use Math to Lie," page 86). Moving
pulses closer together makes them more susceptible to multipath
interference, a problem from which UWB is so far immune. Other
interference issues are likely when several UWB devices are close
together and transmitting at the same time. But even without any
increase, UWB's demonstrated speed already matches the fastest WLANs.

That isn't a fair comparison, because WLANs are available now and UWB is
still theoretical. But we don't have long to wait. Vendors are rushing
to get products out of the labs, and XtremeSpectrum says that it will
ship a chipset aimed at consumer gadgets by the end of June 2002. We
should soon find out whether the hype is justified.

Senior editor Andy Dornan's new edition of his book, The Essential Guide
to Wireless Communications Applications, ISBN 013-0097-187, is published
by Prentice Hall. He can be reached at adornan@ cmp.com.

-

Resources

Ultra-Wideband (UWB) vendor Multispectral Solutions has a useful and (by
vendor standards) unbiased FAQ covering the technology at
www.multispectral.com/uwbfaq.html.

(r)ther Wire & Location has produced an entire CD-ROM filled with PDF
format papers, some technical, covering the history and applications of
UWB. Much of the material is also available online, at
www.aetherwire.com/CDROM/ General/papers.html.

The Berkeley Wireless Research Center, at http://bwrc.eecs.berkeley.edu,
runs the "PicoRadio" project, which aims to develop short-range wireless
date networks with low power consumption.

The FCC's February 14, 2001 decision on UWB, along with PowerPoint
slides and statements from commissioners, can be seen at
www.fcc.gov/Bureaus/Engineering - Technology/News -
Releases/2002/nret0203.html. All of the comments the FCC received from
cellular operators, UWB vendors, Global Positioning System (GPS) users,
and other interested parties have also been posted online, in PDF
format. They are accessible via a Web form at
http://gullfoss2.fcc.gov/prod/ ecfs/comsrch - v2.cgi, which requires the
FCC proceeding number ("98-153").

A good overview of the theory and tests results for and against UWB's
interference with GPS is in the September 2001 issue of GPS World
magazine, posted online at www.gpsworld.com.

For an explanation of how UWB works, see Network Magazine's tutorial
(Lesson 160: Ultra-Wideband Wireless Networks") in the November 2001
issue on page 30. We also have more on GPS, Enhanced-911 (E-911), and
location technology. See "Can M-Commerce Find a Place in Your Network"
on page 38 of the same issue, and "Did Cell Phones Save the White
House?" December 2001 on page 78. All are available online at
www.networkmagazine.com.

http://www.networkmagazine.com.

(c) Copyright 2002 CMP Media LLC.

NETWORK MAGAZINE 06/2002

Signed,
Steven Haigh
President - Melbourne Wireless
www.wireless.org.au

To unsubscribe: send mail to majordomo at wireless.org.au
with "unsubscribe melbwireless" in the body of the message