快猫短视频

鈥淚 WAS thinking about this very deep question,鈥� says Gerard Foschini. 鈥淗ow
much information can you get from one volume of space to another? Then there was
a moment when the equations came together and I thought `ah, that鈥檚 the way to
do it鈥�. That was my Eureka moment.鈥�

Foschini is a mathematician at Bell Labs in New Jersey, and his moment of
inspiration has turned the world of wireless communications on its head. His
brainwave was to exploit the radio reflections and scattered signals that
engineers are usually desperate to avoid. Scattered signals spell trouble. They
cause interference and 鈥済hosting鈥�, as anybody living near a tall building will
tell you. The received wisdom among engineers is that a radio signal should
travel in a straight line and if there are any reflections on the way, somehow
you鈥檝e got to get rid of them.

But Foschini has shown that this traditional mindset is wrong. 鈥淭he equations
tell you that you can use the environment to great effect,鈥� he says. Far from
reducing the quality of signals, reflections can vastly increase the amount of
information that can be sent from one location to another. Use signals reflected
from huge buildings and you can transform mobile phone networks from bottlenecks
in the information superhighway into data fire hoses. Foschini鈥檚 idea is a
revelation for communications engineers. Yet scientists believe that the work
may have far-reaching effects in other areas of science too.

The story began more than 50 years ago when a young scientist called Claude
Shannon, also at Bell Labs, made the first giant strides in understanding the
role that information plays in the world of communications. He defined the
fundamental problem of communication as the task of reproducing at one point in
space a message that was created at another point. Shannon also made clear that,
from an engineer鈥檚 standpoint, the content of such a message is irrelevant. He
went on to work out how reliably such a message could be sent and the
theoretical limit for how much information it could contain. Today, the 鈥淪hannon
limit鈥� is the gold standard that engineers strive for as they design ways to
encode messages for everything from satellite communications systems to mobile
phone networks.

Shannon鈥檚 ideas have survived intact for half a century, but the explosion in
demand for information has forced scientists to come up with new and inventive
ways to send it. One thing in particular has troubled engineers working on
mobile phone networks: what happens when you reach the maximum capacity of the
鈥渃hannel鈥�, the data pipeline between the phone and its local source, the base
station?

The same question has bugged engineers working in other areas of
communication, but usually the answer has been a little more obvious. When one
channel is full, they simply set up another one. So when an optical fibre
becomes saturated, for example, engineers can always lay another, doubling
capacity at a stroke.

In the world of wireless communications, where the radio spectrum has to be
shared, things are not so straightforward. When the available frequencies have
been used up, there is no easy way to increase capacity. You can鈥檛 go to the top
of a hill and add a second or third transmitter in the hope that they will act
as additional pipelines. From a far-away receiver, the transmitters will be
indistinguishable鈥攖heir signals will look like a jumbled transmission from
a single transmitter.

In the mid-1990s, researchers at Bell Labs realised that unless someone came
up with a solution, radio and mobile phone networks were doomed to sit in the
slow lane of the information superhighway. Without more capacity, wireless Web
access, mobile commerce and the prospects for Dick Tracy-style video wristphones
would splutter and stall.

In 1994, Foschini began to look for an answer in Shannon鈥檚 ideas, focusing in
particular on the role of reflected signals. To radio waves, large buildings
such as office blocks and sports stadiums are huge mirrors鈥攕ignals bounce
off them and follow different paths to the receiver. When the direct and
reflected signals arrive together, they interfere, creating the kind of ghost
effect that viewers see when television pictures reflect off tall buildings
nearby. To avoid these reflections, engineers have always tried to ensure that
signals travel directly from transmitter to receiver, and they go out of their
way to place transmitters on hills or on top of tall buildings, hoping to
establish direct line-of-sight links with the receivers.

But Foschini鈥檚 equations told him that this wasn鈥檛 such a good idea. Instead
of avoiding reflections, he realised he could use them to his advantage.
Reflections mean that the signals from the transmitter travel a number of
different paths and arrive at the receiver superimposed. He worked out a way to
process the incoming signals so that if the reflections were more or less
random, it would be possible to reconstruct the original broadcast.

When one transmitter and receiver are involved, this system is no better than
any other, but it comes into its own when an array of transmitters at one point
broadcasts simultaneously to an array of receivers at another. The signals leave
the transmitters and bounce their way to the receivers via a number of
reflectors. By the time they reach the receivers, the signals will be
superimposed in a jumbled mass of noise that, according to conventional
communications theory, might never be unravelled.

But Foschini realised that each signal is scattered differently and follows a
different pathway鈥攁s a result of each transmitter in the array being at a
slightly different location. This separates them slightly in time and space.
Although they arrive at the receivers superimposed, they don鈥檛 look as if they
came from a single transmitter. With some clever processing it should be
possible to separate them.

The key to the technology is a set of high-speed signal processors that look
at the signals from all the receiver antennas simultaneously. First they extract
the strongest signal from the jumble, then they work through the weaker signals
one by one. These are easier to recover once the strongest signal has been
removed, since it is the main source of interference.

Foschini鈥檚 work is a turning point. It means that the amount of information
that can be broadcast through a cluttered space such as a city no longer depends
on a single channel but on the number of transmitters and receivers at either
end. Add another transmitter and receiver and the data rate increases
correspondingly, just like laying another fibre-optic cable. He called the
system BLAST (Bell Labs layered space-time) and began persuading other
scientists to study it further.

Other researchers soon found that Foschini鈥檚 discovery allows them to think
very differently about Shannon鈥檚 ideas. Shannon envisaged a way of communicating
from point to point and devised a coding system that varies in time to achieve
this. But by thinking about the volume of space that the antennas take up,
Foschini has worked out how a certain volume of space can be linked to another
volume by means of communication. Accordingly, his coding system must take into
account time and space. Far from breaking Shannon鈥檚 limit, Foschini has vastly
generalised it.

He and his colleagues soon began trials to test whether the idea would work
in practice. The efficiency of a broadcast system is usually measured by the
number of bits of information per second that can be sent per hertz of radio
bandwidth used up. At best, mobile phone networks currently operate at 5 bits
per second per hertz, yet even with 30 kilohertz of bandwidth, data rates rarely
exceed 50 kilobits per second. By 1998, Foschini had shown that with fewer than
12 antennas at each end, it would be possible to send data up to 20 times
faster. 鈥淭his kind of capacity is simply not possible with conventional
systems,鈥� says Foschini.

Don鈥檛 expect to see BLAST phones in the high street just yet, however. There
are a few obstacles that the team must overcome to make the idea useful in the
real world. For example, so far the work has focused on transmitting from one
array or 鈥渧olume鈥� to another, whereas a mobile phone network involves one base
station transmitting to many phones. Foschini is currently working out how to
extend his ideas to networks.

Built for speed

Then there is the problem of mobility. BLAST was originally developed for
fixed transmitters and receivers, since any motion changes the reflecting
environment and the pathways that the signals takes to the receiver. Lucent is
now working on a mobile version in which the signals are decoded so quickly
that, to the receiver, the moving phone appears stationary. 鈥淲e know that the
system works at highway speeds,鈥� says Bert Hochwald, an electrical engineer at
Lucent who is working on mobile applications for BLAST.

Can you cram enough antennas into a space no bigger than a mobile phone? It
should be possible, the researchers say. Early work on this problem suggests
that if the antennas are placed less than a few centimetres apart, it might even
be possible to transmit at higher data rates. 鈥淭he antennas couple together at
these spacings so the situation is extremely complex,鈥� says Foschini.

And what about those living outside city centres鈥攃ould BLAST still work
in the suburbs and countryside where there are few tall buildings? Foschini is
hopeful. It is possible that engineers could use small, strategically placed
metal reflectors鈥攚hich appear huge to radio waves鈥攊nstead of large
buildings.

Meanwhile, others at the labs have begun to look at the implications of
Foschini鈥檚 work. Theoretical physicist Steve Simon and his colleagues have begun
by studying the link between information and space. 鈥淪uppose you packed a given
volume of space with antennas and started broadcasting, what would happen?鈥� asks
Simon. 鈥淲ould some antennas cancel out while others reinforce, would those on
the surface block transmissions from inside, is there a limit on how much
information you could transmit out of the volume?鈥�

Because of the sheer number of variables involved, the calculations are
hugely complex. But Simon says a clear trend has begun to emerge. In each case
studied, the most important contributions to the signal come from antennas on
the surface of the volume, while those inside tend to be obscured or cancel out.
It seems that surface area, rather than volume, governs the amount of
information that it is possible to extract from any region of space.

These ideas are familiar to at least one other group of scientists. In the
1960s, cosmologists began to explore the properties of black holes. They were
able to work out that black holes have some very basic properties such as mass
and spin, but not much else. Then people began to ask about the relationship
between black holes and entropy, a measure of disorder that is closely linked to
the mathematical definition of information. One of the fundamental laws of
physics鈥攖he second law of thermodynamics鈥攊mplies that the amount of
entropy in the Universe increases inexorably. But if a black hole swallows a
star, for example, what happens to the entropy associated with that matter?
Surely it cannot disappear, since this would contradict the second law?

Jacob Bekenstein, a physicist now with the Racah Institute of Physics at the
Hebrew University in Jerusalem, has worked out an answer. The entropy, he
discovered, is somehow stored on the surface area of a black hole鈥攐n its
event horizon in the jargon of cosmology鈥攁nd this surface area is
proportional to the entropy it has swallowed. The extraordinary implication is
that by thinking about mobile phones, electrical engineers have arrived at a
similar conclusion about the relationship between space and information.

This link could eventually have repercussions for every desktop in the world.
Since computer memories are essentially volumes of space from which information
must be extracted, the finding may place a fundamental limit on the amount that
can be stored in a chip or on a hard disk. 鈥淒oes it place a fundamental limit on
computing memory?鈥� says Simon. 鈥淲e just don鈥檛 know.鈥�

Then there is life itself. Our sense of smell, for example, is a system for
conveying information from the environment to the brain that involves between
500 and 1000 receivers. These receivers are known as receptors: sites in the
nose where smelly molecules bind. When a molecule sticks to a receptor, it
triggers a signal that is sent to the brain for interpretation.

The trouble is that biologists have not been able to decode the rules that
determine how specific molecules or combinations of them create smells, or how
humans are able to distinguish several smells at the same time鈥攅ach of
which might be the result of several molecules. They have long realised that
there cannot be a simple relationship between binding to a particular receptor
or a specific combination of receptors and a smell such as apples or lemon
drops, but are at a loss to explain what is going on.

Maybe something very similar to BLAST, suggests Anirvan Sengupta, a
theoretical physicist at Bell Labs who is studying this question. Sengupta
thinks of the odour molecules as channels along which information is sent, the
receptors as receivers and the brain as the device that carries out signal
processing. He points out that there are many more channels than there are
receivers and that the brain is able to make sense of the smell from many
signals that arrive in parallel. It can even tease apart two or three different
smells present at the same time. The BLAST concept shows that the way to
maximise the amount of information sent is to send it over a multitude of
channels using a number of receivers. 鈥淭here seems to be a definite mathematical
relationship between BLAST and the way smell works,鈥� he says, adding that he is
hoping to collaborate with biologists to explore the idea further.

Tom Schneider, a biologist specialising in information theory at the
Laboratory of Experimental and Computational Biology at the National Cancer
Institute in Frederick, Maryland, points out that many biological systems seem
to have receptors that work in parallel, but in ways we do not yet understand.
鈥淏LAST sounds very promising, but what we lack is good data from these systems
that we can use to test the idea.鈥�

That should change given time. Meanwhile, Foschini and his colleagues are
busy turning BLAST from a neat lab trick to a full-blown commercial reality.
鈥淓very day our understanding increases,鈥� he says. 鈥淚t has tremendous
辫辞迟别苍迟颈补濒.鈥�