U.S. patent application number 09/864241 was filed with the patent office on 2002-04-18 for spread spectrum bit allocation algorithm.
Invention is credited to Shively, Richard Robert, Sonalkar, Ranjan V..
Application Number | 20020044597 09/864241 |
Document ID | / |
Family ID | 21693240 |
Filed Date | 2002-04-18 |
United States Patent
Application |
20020044597 |
Kind Code |
A1 |
Shively, Richard Robert ; et
al. |
April 18, 2002 |
Spread spectrum bit allocation algorithm
Abstract
High transmission capacity in a twisted pair signal line, where
power is limited by a power spectral-density mask and an aggregate
signal power constraint, is obtained by: (1) allocating data to
multitone sub-bands according to a lowest marginal power-cost per
bit scheme and (2) in an environment where an aggregate power
budget remains after all bits have been allocated to all sub-bands
with sufficient margins to carry a bit, assigning additional bits
to sub-bands with otherwise insufficient power margins to carry a
single bit, by frequency-domain-spreading a single bit across
several sub-bands at correspondingly reduced power levels, to
permit the otherwise unacceptable noise levels to be reduced on
average by despreading at the receiving end. Another feature of the
invention, applicable in an environment in which multiple
interfering channels are employed, provides increased signal
throughput by (3) transmitting coherently in a number of multitone
sub-bands, identical blocks of data, with the number of multitone
sub-bands being equal to a number of interfering channels and
multiplying the signal carried by corresponding sub-bands in the
separate interfering channels by a different respective vector from
an orthonormal basis set so that near-end cross-talk is eliminated
upon despreading at the receiving end.
Inventors: |
Shively, Richard Robert;
(Convent Station, NJ) ; Sonalkar, Ranjan V.;
(North Caldwell, NJ) |
Correspondence
Address: |
BANNER & WITCOFF
1001 G STREET N W
SUITE 1100
WASHINGTON
DC
20001
US
|
Family ID: |
21693240 |
Appl. No.: |
09/864241 |
Filed: |
May 25, 2001 |
Related U.S. Patent Documents
|
|
|
|
|
|
Application
Number |
Filing Date |
Patent Number |
|
|
09864241 |
May 25, 2001 |
|
|
|
09527305 |
Mar 16, 2000 |
|
|
|
6285708 |
|
|
|
|
09527305 |
Mar 16, 2000 |
|
|
|
09000842 |
Dec 31, 1997 |
|
|
|
6144696 |
|
|
|
|
Current U.S.
Class: |
375/222 ;
370/344; 375/260 |
Current CPC
Class: |
H04L 5/0044 20130101;
H04L 27/2608 20130101 |
Class at
Publication: |
375/222 ;
375/260; 370/344 |
International
Class: |
H04B 001/38; H04L
027/28; H04B 007/208 |
Claims
1. A transmitting modem receiving digital data from a data source,
modulating carriers to represent said digital data, and applying a
resulting modulated signal to a channel connectable to a receiving
modem, said channel being subject to a power spectral density mask
said transmitting modem comprising: first, second, and third signal
modulators, each with an input; a signal combiner with a combined
output connected to said channel; a serial-to-parallel converter
connected to said data source and to each of said first, second,
and third signal modulator inputs such that said digital data from
said data source is converted to multiple parallel streams applied
respectively to said first, second, and third signal modulators;
each of said first, second, and third signal modulators having a
respective output connected to said signal combiner such that a sum
of output signals of said first, second, and third signal
modulators is applied to said channel; a transfer characteristic of
said channel being such that a first minimum power required to
represent a specified minimum number of bits by modulating in a
first frequency sub-band falls below said power spectral density
mask and such a that a second minimum power required to represent a
second specified minimum number of bits by modulating in each of
second and third frequency sub-bands exceeds said power spectral
density mask; said serial-to-parallel converter being programmed to
feed a first bit of said digital data to said first signal
modulator to represent said first bit by modulating in said first
frequency sub-band at a first power level at least as high as said
first minimum power; said serial-to-parallel converter being
programmed to feed a second bit of said digital data to said second
and third modulators to represent said second bit by coherently
modulating in both said second and said third frequency sub-bands
at a second power level below said first power level, whereby
resulting signals applied in said second and third frequency
sub-bands may be combined by said receiving modem to retrieve said
second bit; said first and second minimum number of bits both being
equal to one in the absence of some other specified constraint.
2. A modem as in claim 1, further comprising fourth and fifth
modulators, said serial-to-parallel converter being programmed to
feed said second bit to said fourth and fifth modulators, whereby
said resulting signals applied in said second through fifth
frequency sub-bands may be coherently combined by said receiving
modem to retrieve said second bit.
3. A modem as in claim 2, wherein said second through fifth
frequency sub-bands are substantially adjacent in frequency such
that they form a continuous spectrum.
4. A modem as in claim 1, wherein said second and third frequency
sub-bands are substantially adjacent in frequency such that they
form a continuous spectrum.
5. A frequency division multiplexor transmitting data from a data
source over a channel, comprising: a signal modulator with an input
and first, second, and third outputs, each transmitting data in a
respective one of first, second, and third frequency bands; a
channel response detector connected to said channel to detect a
transfer characteristic of said channel, said transfer
characteristic including a noise power level and an attenuation of
said channel; a controller connected to said signal modulator to
control an allocation of first and second blocks of data from said
data source for transmission in said first, second, and third
frequency bands; said controller being programmed to transmit said
first block of data in said first frequency band and transmit said
second block redundantly in each of said second and third frequency
bands at a first power level when said channel transfer
characteristic is such that a power level required to transmit said
second block, at a specified bit error rate, in said second
frequency band alone is a first power level and to transmit said
second block in said second frequency band alone when said channel
transfer characteristic is such that a power level required to
transmit said second block, at said specified bit error rate, in
said second frequency band alone at a second power level, said
second power level being higher than said first power level.
6. A frequency division multiplexor as in claim 5, wherein said
controller transmits said second block redundantly in said second
and third frequency bands when an amount of power required to
transmit a single bit in said second frequency band exceeds a power
spectral density mask limit stored by said controller.
7. A frequency division multiplexor as in claim 6, wherein said
controller is programmed to transmit said second data block in said
second frequency band alone when a power required to transmit said
second data block in said second frequency band is less than a
power spectral density mask limit stored by said controller and
greater than an aggregate power limit for all said first, second,
and third frequency bands stored by said controller.
8. A frequency division multiplexor as in claim 5, wherein said
controller is programmed to calculate an array of values each
indicating an amount of power required to transmit a minimum number
of bits in each of said second and third frequency bands and to
sort said array according to said power, said controller being
further programmed to transmit said second block in said second and
third frequency bands responsively to one of said second and said
third frequency bands appearing in said sorted array at a position
indicating that said second bin requires a minimum amount of power
to transmit said minimum number of bits.
9. A frequency division multiplexor as in claim 5, wherein said
controller is programmed to calculate an array of values each
indicating an amount of power required to transmit a minimum number
of bits in each of said second and third frequency bands and to
sort said array according to said power, said controller being
further programmed to transmit said second block in said second and
third frequency bands responsively to said sorted array.
10. A modem, comprising: a frequency-division modulator with a
controller, said modulator transmitting input data in frequency
channels; said controller having a memory storing a power spectral
density (PSD) mask specifying a respective maximum power level
permitted for a signal transmitted in each of said frequency
channels; said controller memory storing an aggregate power limit
specifying a total permitted power for all of said signals; said
controller being programmed to measure and store in said memory a
measured channel transfer characteristic of a communications
channel through which said input data is to be transmitted; said
controller being programmed to transmit respective unique portions
of said input data in said each of said frequency channels,
responsively to said stored aggregate power limit, said PSD mask,
and said measured transfer characteristic being a first transfer
characteristic and to transmit a same portion of said data in at
least two of said frequency channels responsively to said stored
aggregate power limit, said PSD mask, and said measured transfer
characteristic being a second transfer characteristic.
11. A modem as in claim 10, wherein said respective transfer
characteristic is responsive to a specified error rate.
12. A modem as in claim 10, wherein said first transfer
characteristic is characterized by at least one of a lower
aggregate noise level and a lower aggregate attenuation than said
second transfer characteristic.
13. A method for use in a data modulator for allocating bits to
data channel frequencies comprising the steps of: storing mask
power data representing a respective maximum power level for each
of said data channel frequencies; storing aggregate power data
representing a total amount of signal power to be applied in all of
said channel frequencies; allocating bits, such that bits are
allocated up to said respective maximum power level for each of
said channel frequencies and such that each of said bits is
allocated to a single respective one of said channel frequencies;
and when said aggregate power level is not substantially reached in
said step of allocating, further allocating bits to multiples of
said channel frequencies for transmission at reduced power rates
per channel frequency, to permit further bits to be allocated,
until one of said aggregate power limit is substantially reached
and said respective maximum power level is reached.
14. A method as in claim 13 wherein said step of allocating bits is
responsive to a channel transform characteristic sensitive to a
noise power level of said channel and an attenuation of said
channel and includes allocating bits to each channel frequency by
allocating bits to the channel frequency having the lowest marginal
power requirement to transmit an additional bit.
15. A method as in claim 13 wherein said step of further allocating
includes selecting a channel frequency from a among a set of
frequency channels for which no data allocated in said step of
allocating.
16. A method as in claim 15, wherein said step of further
allocating includes selecting a channel frequency from among said
set based on a lowest marginal power requirement for transmitting a
further bit.
17. A method as in claim 13, wherein said step of allocating
includes allocating according to a minimum number of bits per
channel frequency.
18. A method as in claim 13, wherein said step of allocating
includes storing an array of length equal to a number of frequency
bins, the value of each element being a maximum number of bits to
be transmitted.
19. A method as in claim 13, wherein said step of further
allocating includes storing an array of length equal to a number of
frequency bins remaining unallocated in said step of allocating,
the value of each element being an amount of power required to
transmit a minimum number of bits in a frequency corresponding to a
respective one of said frequency bins remaining.
20. Apparatus for allocating bits for data transmission via a
plurality of discrete frequencies comprising tone ordering
circuitry, gain scaling circuitry and an inverse discrete Fourier
transform modulator, the circuitry in combination to: allocate
initial bits to frequencies on a per frequency basis, such that
said initial bits are successively allocated until a maximum power
level for each frequency is at least substantially reached, each of
said initial bits being unique to a given frequency; calculate a
stored total power level for said initial bits allocated to a
plurality of transmit frequencies, and if the stored total power
level is not exceeded, allocate further bits to frequencies for
which no initial bits are allocated, such that each of said further
bits is redundantly allocated to more than one of said
frequencies.
21. Apparatus as in claim 20, wherein said further bits are
allocated sequentially according to the frequency requiring the
least power to transmit an single bit.
22. A frequency-division multiplex (FDM) transmission system for a
channel having multiple subchannels, each of said subchannels being
susceptible to cross-talk interference from another of said
subchannels, said system comprising: a transmitting modem with a
programmable FDM modulator connected to modulate first and second
frequency carriers, representing an input data stream, in each of
first and second subchannels of said channel; a receiving modem
connected to said channel; a modulator programmed to modulate said
first and second frequency carriers coherently to represent a first
subportion of said data stream in said first and second frequency
bands to form first and second signals in said first subchannel;
said modulator being programmed to modulate said first and second
frequency carriers coherently to represent a second subportion of
said data stream in said first and second frequency bands to form
third and fourth signals in said second subchannel; said receiving
modem having a demodulator configured to combine coherently said
first and second signals; said modulator being further programmed
to form said third and fourth signals such that when said
demodulator combines coherently said first and second signals,
cross-talk interference in said first subchannel, caused by
concurrent transmission of said third and fourth signals in said
second channel is diminished in a combined signal resulting
therefrom.
23. A system as in claim 22, wherein said receiving modem combines
said first and second signals such that an incoherent distortion of
said first and second signals in said first subchannel is
attenuated relative to said subportion of said data stream.
24. A system as in claim 22, wherein said modulator is programmed
to weight said first and second signals by a respective one of a
set of codes forming an orthogonal set.
25. A system as in claim 24, wherein said modulator is programmed
to weight said second and third signals by a another one of said
set of codes.
26. A method for reducing near end cross talk comprising the steps
of: forming first and second signals in respective first and second
tones redundantly representing first data to form a first
multi-tone signal such that said first and second signals are
weighted by a first vector of an orthogonal set of codes; applying
said first multi-tone signal to a first interfering channel;
forming third and fourth signals in said respective first and
second tones redundantly representing second data to form a second
multi-tone signal such that said third and fourth signals are
weighted by a second vector of an orthogonal set of codes; applying
said second multi-tone signal to a second interfering channel;
combining said first and second multitone signals such that a
distortion in said first data caused by near end cross talk in
first interfering channel is diminished.
27. A method as in claim 24, wherein said step of combining
includes deweighting said first and second signal and then summing
deweighted versions of said first and second signals.
28. A method as in claim 24, wherein said orthogonal set of codes
are Walsh codes.
29. A method as in claim 24, wherein first and second channels are
twisted pairs of conductors in a metallic telecommunications
network.
30. A frequency division multiplex (FDM) transmission system for
transmitting an input data stream through a channel, comprising: a
transmitting modem with a programmable FDM modulator connected to
modulate first and second frequency carriers, representing an input
data stream, in said channel; a receiving modem connected to said
channel; a modulator programmed to modulate said first and second
frequency carriers coherently to represent a first subportion of
said data stream in said first and second frequency bands to form
first and second signals; said receiving modem having a demodulator
configured to combine coherently said first and second signals to
extract said first subportion such that an incoherent distortion of
said first and second symbols in said first channel is, on average,
reduced in said extracted first subportion.
31. A system as in claim 28, wherein said modulator is programmed
to modulate a third frequency carrier to represent a second
subportion of said data stream, said demodulator being configured
to extract said second subportion without combining said a signal
corresponding to said third frequency carrier with another signal
corresponding to another frequency carrier modulated in said
channel.
32. A method for increasing a data rate in a communication channel
subject to a power spectral density mask, comprising: detecting a
transfer characteristic indicating a required minimum power of a
respective carrier modulated to transmit one bit in each of a
plurality of multitone subchannels of said channel; supplying a
data stream to a modulator; modulating a first set of respective
carriers to represent respective unique portions of said data
stream in at least a subset of those of said multitone subchannels
for which, in said step of detecting indicates said minimum power
falls below a power limit imposed by said power spectral density
mask; modulating a second set of respective carriers to represent
redundantly at least one portion of said data stream in at least a
subset of those of said multitone subchannels for which said step
of detecting indicates said minimum power exceeds a power limit
imposed by said power spectral density mask; receiving said at
least first and second symbols at a receiving end of said
communication channel; combining said at least first and second
symbols received at said receiving end in such a way as to increase
a signal power of said at least first and second symbols and, on
average, reduce incoherent distortion of said at least first and
second symbols; reconstructing said same portion of said data
stream at said receiving end from a combination of said at least
first and second symbols, resulting from said step of
combining.
33. A communications system for communicating data in a channel
having multiple subchannels, comprising: a modulator having an
output for each of said subchannels; and a demodulator having an
input for each of said subchannels; said modulator having an input
connected to receive data from a data source; said modulator being
programmed to modulate separate sets of carriers in each of said
subchannels to represent respective portions of said data; said
modulator being programmed to modulate n separate frequency
carriers coherently in each of said subchannels to represent a one
of said respective portions of said data; n modulated signals
resulting from a modulation of said modulator being output by said
modulator; each of said n modulated signals, upon being received at
said demodulator, including an incoherent component resulting from
attenuation and/or noise in said channel, a first coherent
component resulting from near-end cross talk from said n modulated
signals in the subchannels other than said each, and a second
coherent component output which is the original n modulated signals
output by said modulator; said modulator modulating said n separate
frequency carriers such that when said demodulator demodulates said
n modulated signals upon being received at said demodulator, by
linearly combining said received n modulated signals, in a signal
resulting from said linearly combining, said incoherent component
and said first coherent component are, on average, suppressed and
said second coherent component is, on average, amplified.
34. A transmitting modem receiving digital data from a data source,
modulating carriers to represent said digital data, and applying a
resulting modulated signal to a channel connectable to a receiving
modem, said transmitting modem comprising: first, second, and third
signal modulators, each with an input; a signal combiner with a
combined output connected to said channel; a serial-to-parallel
converter connected to said data source and to each of said first
and second signal modulator inputs such that said digital data from
said data source is converted to multiple parallel streams applied
respectively to said first and second signal modulators; each of
said first and second signal modulators having a respective output
connected to said signal combiner such that a sum of output signals
of said first and second signal modulators is applied to said
channel; said serial-to-parallel converter being programmed to feed
a bit of said digital data to said first and second modulators to
represent said second bit by coherently modulating in both said
first and second frequency sub-bands, whereby resulting signals
applied in said first and second frequency sub-bands may be
coherently linearly combined by said receiving modem to retrieve
said bit and such that incoherent components and coherent but at
least partially orthogonal components of said resulting signals are
attenuated and coherent component of a modulated signal applied by
said first and second modulators is amplified.
35. A method for transmitting data through a channel subject to a
power spectral density mask limit and an aggregate power
constraint, comprising the steps of: detecting a
frequency-dependent transmission characteristic of said channel;
defining fist and second sub-bands of an aggregate transmission
band responsively to a result of said step of detecting, said
aggregate power constraint, and said power spectral density mask
limit; generating a first modulated signal with a first power
dynamic range permitted by said power spectral density mask limit,
said first modulated signal representing said first portion of said
data; generating a second modulated signal with a second power
dynamic range permitted by said power spectral density mask limit,
said second modulated signal representing said second portion of
said data.
36. A method of communicating a signal through a first channel in
which interfering signals are generated by other channels,
comprising the steps of: modulation said signal to generate a first
encoded signal that is orthogonal to signals in said other
channels; and applying said first encoded signal to said first
channel.
37. A method as in claim 34, wherein said first encoded signal is a
CDMA signal.
38. A method as in claim 35, further comprising the step of
applying a second encoded signal to said channel, said second
signal being a CDMA signal with a code orthogonal to a code of said
first encoded signal.
39. A method as in claim 34, wherein said first and said other
channels are adjacent wires forming a cable bundle.
40. A method for communicating information through a cable
containing a plurality of wires, comprising the steps of: dividing
the plurality of wires into groups; and modulating signals that are
to be communicated over different wires within a group with a
modulation technique such that the modulated signal communicated
over any wire in the group is orthogonal to the modulated signal
communicated over any other wire in the group.
41. An apparatus for placement on a customer premises and for
connecting premises communication devices to a wire pair leading to
a communication network, comprising: a network port for connecting
to the wire pair; a set of premises ports, containing at least one
premises port, for connecting premises communication devices to
said apparatus; means for creating a CDM signal by modulating a
source signal supplied by said premises devices with a
code-division-multiplexing code that arranges for the spectrum of
the CDM signal to occupy substantially all of the useful bandwidth
of the wire pair; and means for applying the CDM signal to the wire
pair.
42. The apparatus of claim 41 where the means for applying forms
the primary termination of the wire pair at the customer
premises.
43. The apparatus of claim 41 where the means for applying forms
the sole termination of the wire pair at the customer premises.
44. The apparatus of claim 41 further comprising a TVRC modem that
develops the CDM signal, and a controller for coupling the premises
devices to the TVRC modem.
45. The apparatus of claim 44 where the TVRC modem comprises: a
hybrid circuit connected to the network port, a D/A converter and
an A/D converter connected to the hybrid circuit, an active echo
canceler coupled to the output of the A/D converter, an output
processing unit for outputting digital signals representative of
signals received at the network port, and an input processing unit
for developing signals applied to the D/A converter.
46. The apparatus of claim 44 further including: at least one
premises port in said set of premises ports for connecting ISDN
phones to said apparatus, and an ISDN interface module interposed
between the controller and the port for connecting ISDN phones.
47. The apparatus of claim 44 further including at least one
premises port in said set of premises ports for connecting digital
premises devices to said apparatus, and a digital devices interface
module interposed between the controller and the port for
connecting digital premises devices.
48. The apparatus of claim 44 where the CDM signal comprises a
plurality of modulated tones, and a preselected number of said
tones are devoted by the controller to control functions.
49. The apparatus of claim 48 where at least one of the premises
devices connected to said set of premises ports is asynchronous,
and at least one is synchronous, and where the controller assigns
the tones of the CDM signal to the asynchronous premises devices
and to the synchronous premises devices, as necessary.
50. The apparatus of claim 49 where the controller assigns the
tones of the CDM signal to the asynchronous premises devices and to
the synchronous premises devices as necessary to satisfy the needs
of the asynchronous premises devices.
51. The apparatus of claim 50 where the controller assigns the
tones of the CDM signal dynamically, as the needs change.
52. The apparatus of claim 50 where the controller increases the
tones of the CDM signal that are assigned to the asynchronous
premises devices when an additional asynchronous premises device
becomes active, and reduces the tones of the CDM signal that are
assigned to the asynchronous premises devices when an active
asynchronous premises device ceases to be active.
53. The apparatus of claim 50 where the tones are assigned by the
controller to form groups of tones that are assigned to
asynchronous and synchronous premises devices.
54. The apparatus of claim 44 where at least some of the premises
devices connected to said at least one premises port are
asynchronous, and some are synchronous.
55. The apparatus of claim 54 where the controller includes a
module for directing information of the synchronous premises
devices and the asynchronous premises devices to different
routes.
56. The apparatus of claim 44 where said set of premises ports
includes: at least one analog port, and a POTS interface interposed
between the at least one analog port and the controller.
57. The apparatus of claim 56 where the POTS interface includes a
current sensing circuit coupled to the controller and arranged to
sense the current flowing through the analog port, and where the
controller includes a current computation module that, based on the
sensed current, determines whether a premises device connected to
the analog port is active.
58. The apparatus of claim 57 where the current computation module
determines which of a number of premises devices connected to the
analog port is active.
Description
TECHNICAL FIELD
[0001] This invention relates to discrete multitone transmission
(DMT) of data by digital subscriber loop (DSL) modems and more
specifically to the allocation of bits, respectively, to the
discrete multitones.
BACKGROUND OF THE INVENTION
[0002] In digital communication systems employing multi-channel or
multi-carrier transmission, the most effective allocation of bits
to the channels has been discussed in the literature. The
well-known solution from information theory, analogized to pouring
water over a terrain defined by the noise/attenuation of the
channel transform characteristic, has been found to insure
efficient use of signal power within limits defined by aggregate
power and power spectral density mask limits. However, the method
in some instances may not go as far as possible in exploiting
available power imposed by these limits.
[0003] For heuristic purposes, the prior art and the invention are
discussed in terms of N quadrature amplitude modulation (QAM)
channels with a uniform symbol rate and a non-uniform (unique to
each channel) QAM constellation QAM, a form of combined amplitude
and phase modulation, represents k-bit sets of data by modulating
two (orthogonal) quadrature carriers, cos 2.pi.f.sub.ct and sin
2.pi.ft to generate a pulse whose phase and amplitude convey the
encoded k-bits of information. The QAM signal tone can be viewed as
a phasor in the complex plane, each distinguishable phasor
representing a unique state of the tone identified with one unique
value in a range. Thus, if the channel and signal power are such
that 4 separate phasors can be reliably distinguished, the scheme
allows two bits to be represented. For 3 bits to be represented, 8
phasors must be distinguished and so on. The number of different
phasors or states that are distinguishable in a single tone
(pulse), the QAM constellation, is limited by the signal to noise
ratio of the channel and limits imposed by external standards as
discussed below.
[0004] In a DMT modem, a transmission frequency band is separated
into N sub-bands or frequency bins, each corresponding to one QAM
channel. In a non-ideal channel each sub-band has a different
capacity as a result of the variation of noise and attenuation with
frequency. In addition, external standards impose limits on the
aggregate power of a signal (the power applied in all sub-band
channels) and a cap on the power as a function of frequency defined
by a power spectral density mask.
[0005] The power spectral density mask may be dictated by the
standard used in a particular country implementing the standard
(such as A.N.S.I. standard T1.413-1995). The mask may also be a
design constraint intentionally imposed by a modem designer for
some other reason. For example, a designer may intentionally impose
a constraint that no more than n bits are to be transmitted on
transmit channel frequency. Similarly, the designer may impose a
constraint that a minimum of bits (or no bits) must be transmitted
on a particular tone or frequency. For example, the power limit for
frequencies or tones between 0 and 200 kilohertz must be less than
-40 dBm/Hz (a power level referenced to one milliwatt over 1 Hz
bandwidth). Above 200 kHz (to frequencies in the megahertz of
spectrum), the constraint may be -34 dBm/Hz.
[0006] Referring to FIG. 1, the attenuation+noise characteristics
of a medium can be graphically represented by a floor in a power
spectral graph. The lower curve, the channel transform
characteristic A represents this floor, that is, the combined
effect of noise and attenuation as a function of frequency. A
certain margin of power is required to meet or exceed the minimum
threshold of a signal for reliable data transmission. In other
words, the power of a signal in a given sub-band must be
sufficiently high to carry a minimal (1-bit) QAM tone to obtain a
predefined bit error rate. The minimum margin, that required to
transmit a single bit, is represented by curve B. Curve C
represents the limits imposed by a power spectral density mask
imposed by an eternal communications standard. The other limit is
on the aggregate power, also defined by an external communication
standard, e.g., ANSI Standard T1.413-1995 limits the total power
for all sub-bands to 100 mWatts. Some coding techniques, such as
Wei code described in American National Standard for
Telecommunications--Network and Customer Installation
Interfaces--Asymmetric Digital Subscriber Line Metallic Interface,
ANSI T1.413-1995, may also require a minimum number of bits in a
frequency band if the band is to convey any information at all. In
other words, if the power spectral density mask limit may require
that less energy be used than the minimum required to transmit a
single bit.
[0007] Note that the minimum allowable size of the power margin is,
in part, arbitrary since, to an extent, it is defined in terms of
some a priori rules and technical criteria (which are arbitrary to
the extent that they establish a dividing line between acceptable
and unacceptable error rates; Bit Error Rate or BER) for the given
communication system. Note also that the size of the margin
available for a given sub-band corresponds to the dimension of the
constellation that can be represented in a signal carried in that
QAM channel. That is, the larger the margin in a band, the greater
the number of states that can be reliably distinguished in that
band and the larger the constellation that can be used.
[0008] The above context creates a bit-allocation problem. That is,
given the constraints, how should bits be allocated among the
available QAM channels to provide the highest possible data rates?
DSL modems that use DMT modulation concentrate the transmitted
information in the frequency sub-bands that have the minimum
attenuation and noise. The optimum distribution of transmission
power is obtained by distributing the power according to the
well-known "water pouring" analogy as described in Robert G.
Gallagher, Information Theory and Reliable Communication, John
Wiley and Sons, New York, 1968. Such optimal distribution requires
a strategy for allocating bits to the sub-bands for the idealized
situation where the channel sub-bands approach zero width
(.DELTA.f.fwdarw.0). For discrete bits, the applicable metaphor
could be described as an ice-cube pouring analogy.
[0009] DSL technology was conceived to maximize the throughput on
twisted pair copper wiring with attendant background noise,
time-variant Far End Cross Talk (FEXT) and Near End Cross Talk
NEXT). To determine the transform characteristic of the channel,
the modems negotiate during an initial channel signal-to-noise
ratio (SNR) estimation procedure. During the procedure, the
transmitter sends a known pseudo noise (PN) signal. The receiver
computes the characteristics of the received signal in the form of
a ratio N.sub.k/g.sub.k, where go is the channel gain (inverse of
the attenuation) in frequency band k and N.sub.k is the noise power
in the band k. The literature contains many algorithms for
determining the power distribution across the full frequency
bandwidth for maximum data throughput. As noted above, the optimum
approach for non-uniform Gaussian noise channel divided such that
each band can be considered an additive white Gaussian noise
channel has been proved to be the "water pouring" algorithm of
power distribution. In this case, the g.sub.k/N.sub.k. Profile is
compared to a terrain and the available aggregate power limit to a
fixed supply of water poured over the terrain. The depth of the
water corresponds to the power spectral density. The water pouring
analogy is inappropriate to allocation of power in digital channels
intended for transmission of binary data (bits).
[0010] According to one method of allocating bits (John A C.
Bingham, Multicarrier Modulation for Data Transmission: An Idea
Whose Time Has Come, IEEE Communications Magazine, May 1990,
pp5-14), frequency sub-bands or bins are "filled" with data bits
one bit at a time. A bit is added to the bin for which the marginal
power cost is the lowest. That is, a bit is added to the bin such
that transmission in that bin is the least expensive, relative to
an additional bit in any other bin, in terms of power needed for
the resulting signal constellation to be received at a predefined
BER The filling procedure is followed until the total Power Budget
is used up. Since power can only be allocated in discrete amounts
corresponding to each bit, the procedure is likened, as mentioned,
to an ice-cube filling procedure rather than a water-filling
procedure.
OBJECTS AND SUMMARY OF THE INVENTION
[0011] It is an object of the invention to provide a method for
transmission in a multitone communication system together with an
algorithm for allocating bits in the system.
[0012] It is an object of the invention to provide a method for
transmission in a multitone communication system subject to an
aggregate signal power constraint together with an algorithm for
allocating bits in the system.
[0013] It is an object of the invention to provide a method for
transmission in a multitone communication system subject to a
signal-power spectral density mask constraint together with an
algorithm for allocating bits in the system.
[0014] It is an object of the invention to provide a method for
transmission in a multitone communication system subject to an
aggregate signal power constraint and a signal-power spectral
density mask constraint together with an algorithm for allocating
bits in the system.
[0015] It is another object of the invention to provide a method
for transmitting data over multiple interfering channels.
[0016] It is another object of the invention to provide a method
for transmitting data over multiple interfering channels and a
method for reducing interference between the interfering
channels.
[0017] It is another object of the invention to reduce near end
cross talk between DSL modems communicating over the same
cable.
[0018] Briefly, high transmission capacity in a twisted pair signal
line, where power is limited by a power spectral-density mask and
an aggregate signal power constraint, is obtained by: (1)
allocating data to multitone sub-bands according to a lowest
marginal power-cost per bit scheme and (2) in an environment where
an aggregate power budget remains after all bits have been
allocated to all sub-bands with sufficient margins to carry at
least one bit, assigning additional bits to sub-bands with
otherwise insufficient power margins to carry a single bit, by
frequency-domain-spreading a single bit across several sub-bands at
correspondingly reduced power levels, to permit the otherwise
unacceptable noise levels to be reduced on average by despreading
at the receiving end. In an environment in which multiple
interfering channels are employed, signal throughput is increased
by (3) forming a number of sub-bands for spreading blocks of data
that is equal to a number of interfering channels and multiplying
the signal carried by corresponding sub-bands in the separate
interfering channels by a different respective vector from an
orthonormal basis set so that near-end cross-talk is eliminated
upon despreading at the receiving end.
[0019] Note that "spreading" as used in the present application,
refers to a process applied at a stage where the signal is
decomposed into spectral elements, so that it can be applied
selectively to frequency components, in contrast to conventional
spreading found in, for example, wireless (cellular) telephony,
where spreading is applied to the signal time series, and affects
(spreads) all elements of the spectrum equally as a
consequence.
[0020] According to the invention bit allocation may be performed
to optimize throughput within aggregate power and power spectral
density mask limits. Some method, such as the approach identified
above with the water pouring analogy, may be used for this bit
allocation. The process of bit allocation will be limited either by
the mask limit or the aggregate signal power limit. If after
efficient allocation, the total signal power is less than the
aggregate power limit, there will usually be unused sub-bands.
These unused sub-bands were rejected in the initial bit-allocation
process because the available power margin in them was insufficient
to transmit a single bit. That is, the channels were identified as
unusable because transmitting a single bit was found to exceed the
mask limit for the channel. In this case, where the bit allocation
process is limited by the mask, the channels with low power margins
are used to transmit information by spreading a single block of
data (one or more bits) over multiple channels and then despreading
them at the receiver.
[0021] The device of spreading and despreading over multiple
channels also provides a mechanism for reducing near end cross talk
(NEXT). The context to which this device applies is a packet
consisting of I interfering channels and nI carriers in each
channel. For example, the channels could be four wire pairs, in
each, some multiple of four carriers are used to convey information
by spreading a single block over each of four carriers to transmit,
and then despreading at the receiver. At the transmitter, however,
the signals in each interfering channel are multiplied by one
element of an I-dimensional orthogonal code (such as a binary
code). At the receiving end, the signals are multiplied again by
the respective opposite orthogonal code and then despread. The
process of despreading not only reduces incoherent noise as in the
embodiment discussed above, but it also substantially eliminates
NEXT because the interference generated in all the frequency
channels, being derived from orthogonal set, cancel each other.
Thus in a channel of four twisted pairs of wires, each pair
transmits a different block of data but every different block is
spread over four carriers in a given wire pair. The signal
transmitted over each of the four wire pairs is assigned one of
four orthogonal codes. Summing each block spread over the four
frequency channels causes mutual cancellation of the four induced
cross-talk signals of the four wires that were multiplied by the
four orthogonal codes.
[0022] Discrete Multitone (DMT) modulation serves as a framework to
demonstrate the spreading process. An input data stream is
segmented into small blocks of bits, and each such block is
re-expressed as a complex number. For example, a constellation of
16 possible discrete complex number values can be used to convey 4
bits, since 16 different states are required to represent 4 bits.
The resultant array of complex numbers is inverse-Fourier
transformed to synthesize a time series, Y(t), that represents a
sum of multiple distinct sinusoids. (A complex conjugate array of
complex numbers is used as an input to the Inverse Fast Fourier
Transform process to assure a real resultant time series.)
[0023] Each of the complex numbers used to encode data therefore
plays the role of a complex spectral coefficient. That is, each
defines the amplitude and phase of one of the orthogonal sinusoids
included in the transmitted waveform. The number of discrete points
in the constellation for each of the bands is a consequence of the
measured attenuation and noise level in that frequency band, based
on a bit-allocation process that need not be described here.
[0024] In both of the above schemes, the signal power in each
frequency carrier is reduced in proportion to the number of
carriers used. Also, in both schemes, the information relating to
the number of bits per block, the frequency channels over which
blocks are to be spread, etc. must be shared between the
transmitter and the receiver. Regarding the latter scheme, the
transmitter and receiver must also share the orthogonal codes to be
used for each twisted pair, though these can be established on a
permanent basis.
[0025] According an embodiment, the invention provides a
transmitting modem that receives digital data from a data source
and modulates separate carriers to represent the digital data. The
modulated signal is applied to a channel connected to a receiving
modem. The channel is subject to a power spectral density mask. The
transmitting modem includes first, second, and third signal
modulators, each with an input. The modem also has a signal
combiner with a combined output connected to the channel and a
serial-to-parallel converter connected to the data source and to
each of the first second, and third signal modulator inputs. The
connection is such that the digital data from the data source is
converted to multiple parallel streams applied respectively to the
first, second, and third signal modulators. Each of the first,
second, and third signal modulators has a respective output
connected to the signal combiner such that a sum of output signals
of the first, second, and third signal modulators is applied to the
channel. A transfer characteristic of the channel is such that a
first minimum power required to represent a specified minimum
number of bits by modulating in a first frequency sub-band falls
below the power spectral density mask and such a that a second
minim power required to represent a second specified minimum number
of bits by modulating in each of second and third frequency
sub-bands exceeds the power spectral density mask. The
serial-to-parallel converter is programmed to feed a first bit of
the digital data to the first signal modulator to represent the
first bit by modulating in the first frequency sub-band at a first
power level at least as high as the first minimum power. The
serial-to-parallel converter is also programmed to feed a second
bit of the digital data to the second and third modulators to
represent the second bit by coherently modulating in both the
second and the third frequency sub-bands at a second power level
below the first power level, whereby resulting signals applied in
the second and third frequency sub-bands may be combined by the
receiving modem to retrieve the second bit. The first and second
minimum number of bits are both equal to one in the absence of some
other specified constraint.
[0026] According another embodiment, the invention provides a
frequency division multiplexor transmitting data from a data source
over a channel. The multiplexor has a signal modulator with an
input and first, second, and third outputs, each output
transmitting data in a respective one of first, second, and third
frequency bands. A channel response detector connected to the
channel detects a transfer characteristic of the channel, the
transfer characteristic including a noise power level and an
attenuation of the channel. A controller connected to the signal
modulator controls an allocation of first and second blocks of data
from the data source for transmission in the first, second, and
third frequency bands. The controller is programmed to transmit the
first block of data in the first frequency band and transmit the
second block redundantly in each of the second and third frequency
bands at a first power level when the channel transfer
characteristic is such that a power level required to transmit the
second block, at a specified bit error rate, in the second
frequency band alone is a first power level. However, when the
channel transfer characteristic is such that a power level required
to transmit the second block, at the specified bit error rate, in
the second frequency band alone at a second power level, the second
power level being higher than the first power level, the controller
transmits the second block in the second frequency band alone.
[0027] According still another embodiment, the invention provides a
modem with a frequency-division modulator and a controller. The
modulator transmits input data in separate frequency channels. The
controller has a memory that stores a power spectral density (PSD)
mask specifying the maximum power levels permitted for each of the
frequency channels. The controller's memory also stores an
aggregate power limit specifying a total permitted power for all of
the signals in all of the channels. The controller is programmed to
measure and store in the memory the channel transfer characteristic
of a communications channel through which the input data is to be
transmitted. The controller is also programmed to transmit
respective unique portions of the input data in of the frequency
channels based on the stored aggregate power limit, the PSD mask,
when the measured transfer characteristic is a first transfer
characteristic. The controller is programmed to transmit a same
portion of the data in at least two of the frequency channels
responsively to the stored aggregate power limit, the PSD mask when
the measured transfer characteristic is a different transfer
characteristic.
[0028] According still another embodiment, the invention provides a
method for use in a data modulator for allocating bits to data
channel frequencies. The method includes the following steps. (1)
storing mask power data representing a respective maximum power
level for each of the data channel frequencies; (2) storing
aggregate power data representing a total amount of signal power to
be applied in all of the channel frequencies; (3) allocating bits
on a per frequency basis, such that bits are successively allocated
until the respective maximum power level is at least substantially
reached for each of the channel frequencies and such that each of
the bits is allocated to a single respective one of the channel
frequencies; and (4) when the aggregate power level is not
substantially reached in the step of allocating, further allocating
bits to multiples of the channel frequencies for transmission at
reduced power rates per channel frequency, to permit further bits
to be allocated, until one of the aggregate power limit is
substantially reached and the respective maximum power level is
reached for each of the data channel frequencies.
[0029] According still another embodiment, the invention provides
an apparatus that allocates bits for data transmission via a
multiple discrete frequencies. The apparatus has tone ordering
circuitry, gain scaling circuitry and an inverse discrete Fourier
transform modulator. The circuitry in combination allocates initial
bits to frequencies on a per frequency basis, such that the initial
bits are successively allocated until a maximum power level for
each frequency is at least substantially reached, each of the
initial bits being unique to a given frequency. The circuitry also
calculates a stored total power level for the initial bits
allocated to a plurality of transmit frequencies, and if the
stored, total power level is not exceeded, allocate further bits to
frequencies for which no initial bits are allocated, such that each
of the further bits is redundantly allocated to more than one of
the frequencies.
[0030] According another embodiment, the invention provides a
frequency-division multiplex (FDM) transmission system for a
channel having multiple subchannels, each of the subchannels being
susceptible to cross-talk interference from another of the
subchannels. The system comprises a transmitting modem with a
programmable FDM modulator connected to modulate first and second
frequency carriers, representing an input data stream, in each of
first and second subchannels of the channel. Also, the system
includes a receiving modem connected to the channel and a modulator
programmed to modulate the first and second frequency carriers
coherently to represent a first subportion of the data stream in
the first and second frequency bands to form first and second
signals in the first subchannel. The modulator is programmed to
modulate the first and second frequency carriers coherently to
represent a second subportion of the data stream in the first and
second frequency bands to form third and fourth signals in the
second subchannel. The receiving modem has a demodulator configured
to combine coherently the first and second signals. The modulator
is also programmed to form the third and fourth signals such that
when the demodulator combines coherently the first and second
signals, cross-talk interference in the first subchannel, caused by
concurrent transmission of the third and fourth signals in the
second channel, is diminished in a combined signal resulting
therefrom.
[0031] According to still another embodiment, the invention
provides a method for reducing near end cross talk. The method
performs the following steps. (1) forming first and second signals
in respective first and second tones redundantly representing first
data to form a first multi-tone signal such that the first and
second signals are weighted by a first vector of an orthogonal set
of codes; (2) applying the first multi-tone signal to a first
interfering channel; (3) forming third and fourth signals in the
respective first and second tones redundantly representing second
data to form a second multi-tone signal such that the third and
fourth signals are weighted by a second vector of an orthogonal set
of codes; (4) applying the second multi-tone signal to a second
interfering channel; and (5) combining the first and second
multitone signals such that a distortion in the first data caused
by near end cross talk in first interfering channel is
diminished.
[0032] According to still another embodiment, the invention
provides a frequency-division multiplex (FDM) transmission system
for transmitting an input data stream through a channel. The system
has a transmitting modem with a programmable FDM modulator
connected to modulate first and second frequency carriers,
representing an input data stream, in the channel. The system also
uses a receiving modem connected to the channel. A modulator is
programmed to modulate the first and second frequency carriers
coherently to represent a first subportion of the data stream in
the first and second frequency bands to form first and second
signals. The receiving modem has a demodulator configured to
combine coherently the first and second signals to extract the
first subportion such that an incoherent distortion of the first
and second symbols in the first channel is, on average, reduced in
the extracted first subportion.
[0033] According to still another embodiment, the invention
provides a method for increasing a data rate in a communication
channel subject to a power spectral density mask. The method
follows these steps: (1) detecting a transfer characteristic
indicating a required minimum power of a respective carrier
modulated to transmit one bit in each of a plurality of multitone
subchannels of the channel; (2) supplying a data stream to a
modulator, (3) modulating a first set of respective carriers to
represent respective unique portions of the data stream in at least
a subset of those of the multitone subchannels for which, in the
step of detecting indicates the minimum power falls below a power
limit imposed by the power spectral density mask; (4) modulating a
second set of respective carriers to represent redundantly at least
one portion of the data stream in at least a subset of those of the
multitone subchannels for which the step of detecting indicates the
minimum power exceeds a power limit imposed by the power spectral
density mask; (5) receiving the at least first and second symbols
at a receiving end of the communication channel; (6) combining the
at least first and second symbols received at the receiving end in
such a way as to increase a signal power of the at least first and
second symbols and, on average, reduce incoherent distortion of the
at least first and second symbols; (6) reconstructing the same
portion of the data stream at the receiving end from a combination
of the at least first and second symbols, resulting from the step
of combining.
[0034] According still another embodiment, the invention provides a
communications system for communicating data in a channel having
multiple subchannels. The system has a modulator with an output for
each of the subchannels. It also employs a demodulator with an
input for each of the subchannels. The modulator has an input
connected to receive data from a data source. The modulator is
programmed to modulate separate sets of carriers in each of the
subchannels to represent respective portions of the data. The
modulator is also programmed to modulate n separate frequency
(resulting from a modulation of the modulator being output by the
modulator) carriers coherently in each of the subchannels to
represent a one of the respective portions of the data. Each of the
n modulated signals, upon being received at the demodulator,
includes an incoherent component resulting from attenuation and/or
noise in the channel, a first coherent component resulting from
near-end cross talk from the n modulated signals in the subchannels
other than the each, and a second coherent component output which
is the original n modulated signals output by the modulator. The
modulator modulates the n separate frequency carriers such that
when the demodulator demodulates the n modulated signals, upon
being received at the demodulator, by linearly combining the
received n modulated signals, in a signal resulting from the
linearly combining, the incoherent component and the first coherent
component are, on average, suppressed and the second coherent
component is, on average, amplified.
[0035] According still another embodiment, the invention provides a
transmitting modem receiving digital data from a data source,
modulating carriers to represent the digital data, and applying a
resulting modulated signal to a channel connectable to a receiving
modem. The transmitting modem has first, second, and third signal
modulators, each with an input. The modem also has a signal
combiner with a combined output connected to the channel and a
serial-to-parallel converter connected to the data source and to
each of the first and second signal modulator inputs such that the
digital data from the data source is converted to multiple parallel
streams applied respectively to the first and second signal
modulators. Each of the first and second signal modulators has a
respective output connected to the signal combiner such that a sum
of output signals of the first and second signal modulators is
applied to the channel. The serial-to-parallel converter is
programmed to feed a bit of the digital data to the first and
second modulators to represent the second bit by coherently
modulating in both the first and second frequency sub-bands,
whereby resulting signals applied in the first and second frequency
sub-bands may be coherently linearly combined by the receiving
modem to retrieve the bit and such that incoherent components and
coherent but at least partially orthogonal components of the
resulting signals are attenuated and coherent component of a
modulated signal applied by the first and second modulators is
amplified.
[0036] According to still another embodiment, the invention
provides a method for transmitting data through a channel subject
to a power spectral density mask limit and an aggregate power
constraint. According to the method, a frequency-dependent
transmission characteristic of the channel is detected. First and
second sub-bands of an aggregate transmission band are defined
responsively to a result of the step of detecting, the aggregate
power constraint, and the power spectral density mask limit. A
first modulated signal is generated, the signal having a first
power dynamic range permitted by the power spectral density mask
limit, the first modulated signal representing the first portion of
the data. A second modulated signal is generated, the second signal
having a second power dynamic range permitted by the power spectral
density mask limit, the second modulated signal representing the
second portion of the data.
BRIEF DESCRIPTION OF THE DRAWINGS
[0037] In the drawing,
[0038] FIG. 1 shows an arbitrary transform characteristic of an
arbitrary channel with multitone channels, a power spectral density
mask limit, and a minimum power required to transmit a single bit,
assuming a specified error and symbol rates, superimposed
thereon;
[0039] FIG. 2 shows modems in communication over one or more
twisted wire pairs for purposes of describing an embodiment of the
invention;
[0040] FIG. 3 shows a general diagram of elements of a
communication system for purposes of describing an embodiment of
the invention;
[0041] FIG. 4 shows a digital modulator and demodulator connected
by a communication channel for a multitone QAM system for purposes
of describing an embodiment of the invention;
[0042] FIGS. 5 & 6 show a flow diagram for purposes of
describing a control method and apparatus controlling the
embodiment of FIG. 4.
[0043] FIG. 7 shows a conceptual rendering of how signals in
separate channels may interfere and how the resulting distortion
may be canceled according to an embodiment of the invention.
DESCRIPTION OF THE PREFERRED EMBODIMENT
[0044] Referring to FIGS. 1 and 2, a transmitting modem 31 is
connected to a receiving modem 32 by a cable 33 having four twisted
pairs of conductors. In long loop systems where cable 3 is of
length of the order 18,000 feet or more, high signal attenuation at
higher frequencies (greater than 500 kHz) is usually observed. This
characteristic of cable 33 is represented graphically by curve A in
FIG. 1.
[0045] For convenience of description, the details of digital
modulator 14 and digital demodulator 16 are described in terms of a
QAM multitone system, although the invention is applicable to other
kinds of multi-carrier and multi-channel signaling as will be
understood by those skilled in the art in light of the teachings
disclosed herein.
[0046] Referring now also to FIG. 3, Modems 31 and 32 contain a
source encoder 11, a channel decoder 12, a digital modulator 14, to
take in and transmit data from a data source 11. Modems 31 and 32
also contain a digital demodulator 16, a channel decoder 17, and a
source decoder 18 to receive the data and supply it to a data sink
19.
[0047] As will be recognized by those skilled in the art, source
encoder 12 compresses data from data source 11 and applies the
result to channel decoder 13 which error correction/detection data
and applies the result to digital modulator 14. Digital modulator
acts as the interface with the communication channel by modulating
the data to generate a signal that can be applied to the
communication channel 15. Digital demodulator 16 constructs a data
stream from the modulated signal and applies it to channel decoder
17. Channel decoder 17 corrects errors in the data stream and
applies the corrected data to source decoder 18 which decompresses
the data and outputs the decompressed data to data sink 19.
[0048] Referring to FIG. 4, In a QAM multitone modulation, the
spectrum is broken into multiple sub-bands or QAM channels. Digital
modulator 14 generates NQAM signal tones, one for each QAM channel.
Each i.sup.th QAM channel carries k.sub.i bits of data. A
serial-to-parallel buffer 41 segments a serial stream of digital
data into N frames, each having allocated to it k.sub.i bits of
data. These are applied to respective inputs of a multicarrier
modulator 42 which generates a QAM tone for each channel.
Multi-carrier modulator generates N QAM tones, one for each
channel, at the same symbol rate but with a respective
constellation for each channel. That is, the i.sup.th QAM channel
carries an 2.sup.ki-ary QAM tone, a tone with 2.sup.ki signal
points. Multi-carrier modulator modulates N subcarriers by
corresponding symbols to generate the N QAM signal tones using an
inverse digital Fourier transform. The allocation of bits in
serial-to-parallel buffer 41 is discussed in detail below.
[0049] A parallel-to-serial converter 43 adds a cyclic prefix (one
known method of preventing intersymbol interference) and sums the
separate modulated data and passes the resulting data stream
through an D/A converter 44 yielding a single analog signal stream.
After the analog signal reaches receiving modem 32, the opposite
operation occurs in A/D converter 46, serial-to-parallel converter
47, and multicarrier demodulator 48 and detector 49. Multicarrier
demodulator 48 strips the modulating signal from the carrier, that
is, it converts the QAM tone data into data representing the
original modulating symbols. Detector 49 maps the resulting symbols
into a set of bits either by quantizing or soft-decision
quantization. These symbols are then queued up in a serial data
stream by parallel to serial converter 50.
[0050] Referring again to FIG. 1, the allocation of bits in
serial-to-parallel buffer 41 is performed according to an algorithm
which is now described. In FIG. 1, the channel transform
characteristic (an attenuation/noise floor) A, in the range of
sub-band b1, is such that at least one bit can be represented by a
QAM-encoded tone signal transmitted in that sub-band. The same is
true for sub-band b2. However, the amount of power available in
sub-bands b3 and b4 are such that a QAM tone generated within the
power ceiling overlying those bands would not meet the BER and
symbol rate required (the minimum power to represent 1 bit being
represented by curve C).
[0051] If, in the process of assigning bits to the available
frequency sub-bands or "bins", the total aggregate power limit is
not exceeded, then there may be otherwise "wasted" power available
due to the inability to put signals in frequency sub-bands for
which the spectral density mask is too close to the
attenuation/noise floor A for a single bit to be transmitted but
for which there is still available "room" for signal power. Such
noisy and/or highly attenuated sub-bands can occur for example in
long-run twisted pair conductors. Unutilized but otherwise
available power could be put in the frequency sub-bands such as b3
and b4 for which the noise-floor (curve A) is too high to permit a
single bit to be properly represented. According to the invention,
digital modulator 14 replicates ("spreads") a k-bit symbol over
multiple adjacent bands with correspondingly less energy in each
band. At the receiving end, detector 49 coherently recombines
("despreads") the redundant symbols in the noisy/attenuated
sub-bands. In recombining the symbols, the symbols are simply
arithmetically added. Because the noise is incoherent while the
signal is coherent, the noise tends to be averaged out while the
signal is reinforced by the addition process.
[0052] In view of the above process of spreading and despreading
certain channels according to the allocation scheme defined above,
detector 49 is capable not only of detecting signal transmitted
conventionally in one frequency bin, but, also of selectively
summing the blocks previously spread and transmitted over the
noisy/attenuated sub-bands. Of course this is done under
programmable control so that, depending on the results of the
initial channel SNR estimation procedure, command signals are sent
to both the transmitting and receiving modems to appropriately
perform the spreading and despreading processes.
[0053] In a practical implementation of the bin-filling procedure
described by Bingham (see Background section), additional
constraints such as minimum and maximum number of bits per bin, a
power spectrum mask, and a desired set of bit rates may also be
considered. Within these constraints and those of the power
spectral density mask and the aggregate power limit discussed
above, the bin-filling procedure would be followed until the
process is halted by any of these limits. Then a second process of
allocation is performed to assign bits to frequency channels that
are unallocated due to their not having sufficient
power-per-channel to transmit the minimum number of bits. The
second procedure determines if these channels can be used within
the constraints imposed by the other limits. Thus, a first step
applies a first algorithm to allocate bits to each channel within
the constraints of aggregate power, maximum power per channel (PSD
mask), and the maximum and minimum number of bits per channel. Note
that any algorithm that allocates one constellation to each
frequency bin would constitute an appropriate first step. A second
algorithm is applied to allocate identical blocks of data to
multiple channels through spreading as discussed above. The second
algorithm would allocate bits to channels if sufficient power
margin is available in multiple channels in the aggregate to
transmit an additional minimum-sized block of data without
exceeding the aggregate power limit. However, it may turn out that
transmission of some blocks through spreading is cheaper in terms
of power use than transmitting those as allocated during the first
allocation procedure. In that case, an approach that employed
spread and unspread signaling in an approach that allocated each
bit to the channel or channels associated with the least marginal
power consumption. would be an alternative approach that might lead
to more channels being used for spreading than for the two-step
approach discussed in detail below. Generally, the most efficient
utilization of a channel is to allocate data to channels with the
highest SNR, which is the whole idea behind the water-pouring
analogy, so the latter approach is not preferred. However, there
may be channels with certain kinds of transform characteristics for
which such an approach would produce improved performance.
[0054] Referring to FIG. 5, a practical bit-allocating algorithm is
proposed in a U.S. Patent Application entitled Method and Apparatus
for Allocating Data for Transmission via Discrete Multiple
Tones--Sonalkar, et. al. According to this procedure, bit
allocation is performed with consideration to all practical
constraints discussed above. In the method described in the above
reference, the entirety of which is incorporated herein by
reference, bits are allocated according to the requirements of an
aggregate power constraint, a power-spectral density mask limit,
and minimum and maximum bits per symbol (e.g., QAM tone).
[0055] According to an invention described by Sonalkar, et. al. in
the application identified above, the following constraints are
considered first: the maximum number of bits allowed in each tone
or frequency bin for transmission and the maximum power to be
transmitted in each bin (the power-spectral-density mask). Then,
the aggregate power constraint (power budget) is applied in a
second process. During the second step, if necessary, bits are
removed according to a bit removal procedure that removes the bit
from the channel, the removal of which results in the greatest
decrease in aggregate power required. With this process, the fewest
bits are removed because the bit that produces the highest marginal
gain in power savings is selected for removal (deallocation).
[0056] In the first procedure, the maximum number of bits per bin
is calculated first with consideration only to the power mask
constraint and any minimum and maximum bits per bin (channel)
constraints. First, in step S21, the channel's transform
characteristic is calculated during an initialization period. Then,
in step S22, the maximum number of bits that can be allocated to
each bin is calculated. This computation is done from a
relationship between the number of bits (b.sub.k) to be transmitted
in a bin and the power needed to transmit that many bits (b.sub.k)
in bin number k For purposes of this calculation, the PSD
mask-limited value for bin k serves as the power available for
transmission in the channel corresponding to bin k.
[0057] At step S23, the total power requirement over all channels
is calculated over all the transmit frequency bins. Then in step
S24, the total power requirement is compared to the aggregate power
limit. From the standpoint of the aggregate power limit, for
example 100 milliwatts total maximum power, the bins may be
considered to be overflowing at this point. If the total maximum
power is not exceeded, then, the first procedure is completed and a
second procedure where the spreading techniques apply, is
implemented (See FIG. 5). On the other hand, if the total power
limit is exceeded, then the first procedure continues with a bit
removal process, in step S25 as follows.
[0058] The bit removal process is implemented to reduce the bit
allocation so that the total power required for transmission meets
the aggregate power limit. According to a preferred bit removal
process, bits are removed sequentially, each bit being selected in
turn as the bit, the removal of which, produces the greatest
marginal gain in terms of recovered power. The first bit removed
then is the bit that cost the most power to transmit. After the bit
is deallocated, the corresponding power saved is subtracted from
the total power. If the total power constraint is still exceeded
after removal of the bit, the next most power-consuming bit is
removed. This process is followed iteratively until the power
constraint is no longer exceeded. Once the aggregate power limit is
met, the first procedure is completed and no second procedure
follows.
[0059] Referring now also to FIG. 5, in preparation for the second
procedure, the first procedure sketched above is performed and it
is confirmed that residual power remains. That is, that the
difference between the aggregate power limit and the sum of the
signal power over all frequencies subject to the other limits
(minimum and maximum bits per channel and PSD mask limit) is
greater than zero. Note that if the first procedure of FIG. 4 gets
to the second phase of step S25, there will be no residual power
and the procedure of FIG. 5 will not be performed.
[0060] The decibel sum of attenuation plus noise power, plotted as
a function of frequency, may be perceived as a "terrain" on which
the communication channel operates. A classic algorithm called
"water filling" refers to continually budgeting signal power, which
is the constrained resource, at the low points of this terrain
first to achieve the optimum efficiency in terms of information per
unit of power. One approach to multi-band assignment, therefore, is
to continue this process by selecting the next four bands from a
list of bands sorted with respect to the attenuation-plus-noise
parameter or the power required to transmit a block of data. While
this guarantees the most efficient use of the remaining spectrum
under the set of constraints described above, such a procedure does
not in general yield 4-tuples that are adjacent in frequency. This
procedure is adopted in the process described next.
[0061] The second procedure begins at step S2. An index is set to 1
and a number of bins for spreading (the number of channels over
which each block will be spread). This can be done as part of an
optimization procedure or selected a priori. To limit the
administrative and communication overhead associated with
initialization and bit assignment (as well as the computational
complexity to explore combinatorics of a large number of options),
there is strong motivation to limit the number of grouping options
to a single option. That is, either bits are assigned to individual
bands as in the prior art, or where spreading is applied, a single
sized m-tuple of frequency bands is allowed. (Choosing a larger
value of m consumes frequency bandwidth more rapidly, choosing a
smaller value limits the SNR deficit that can be overcome.)
Empirical studies of actual telephone loop propagation
characteristics indicates the payoff is greatest for making the
number of bins, m, equal to 4 under this set of conditions.
[0062] Net, at step S3, an array is calculated containing the power
required to transmit one bit for all bins that were not allocated
in the procedure of FIG. 5. The array indicates the bin by a
frequency pointer (a label), the number of bits already assigned to
that bin, and the power requirement to add another bit to that bin.
Now, the power required to transmit b.sub.k bits is given by: 1 E k
= ( 2 b k - 1 ) ( K N k 3 g k G c ) , ( 1 )
[0063] where:
[0064] b.sub.k is the number of bits carried in frequency bin k,
E.sub.k, is the power required in bin k to transmit the b.sub.k
bits, g.sub.k/N.sub.k is the measured gain to noise ratio in bin k,
and G.sub.c is the coding gain. K is given by: 2 K = [ Q - 1 ( P e
N e ) ] 2 , ( 2 )
[0065] where P.sub.e is the error probability given by: 3 P e = 4 (
1 - 1 2 n i ) Q [ a 2 N 0 ] , ( 3 )
[0066] the Q function is the partial integral of the normal
distribution function given by: 4 Q ( x ) = x .infin. 1 2 e - u 2 2
u , ( 4 )
[0067] N.sub.e is a number that assumes a value in the range of 2
to 4, in the preferred embodiment of the invention, N.sub.e is
equal to 3. Since the second procedure for allocating bits to the
impaired channels (the channels for which the SNR ratio does not
permit a bit to be transmitted without hitting the PSD mask limit)
allocates one bit at a time, b.sub.k can be set to 1 and equation
(1) thus reduces to (also dividing by 4 to account for the
spreading): 5 E k = [ 1 4 ] K N k 3 g k G c ( 5 )
[0068] Thus in calculating the power bin array, the number of bits
alma allocated to each of the bins is taken into account in
calculating the power requirement. So the power requirement
calculation is a marginal power required to add another bit to the
bin. In calculating the amount of power required, if the calculated
power would exceed the PSD mask limit for that bin, a very large
number is used for the energy requirement so that when the array is
sorted in order of increasing power, these bins with low PSD limits
end up at the bottom of the sorted list.
[0069] In step S4, the power bin array is sorted in ascending order
of power requirement so that the element with an index of one
corresponds to the bin with the lowest energy requirement. Once the
array is sorted, the array is repopulated with the bins with power
requirements below the large number (used to put bins for which an
additional bit would exceed the PSD mask limit) and a
bins-remaining-counter, indicating the number of bins to which a
bit can still be assigned, is decremented to reflect the current
number of bins now available.
[0070] In step S7, the bins-remaining counter is checked to see if
m-1 remain. If insufficient bins remain to spread a bit, the
procedure terminates.
[0071] In step S5, the residual power (difference between aggregate
power requirement and the power required to transmit all assigned
bits) is calculated from all assigned bins using equation (1).
Next, in step 9, the total power to transmit one bit in the
adjacent m bins is calculated. Note that the m adjacent bins are
adjacent in terms of the index value corresponding to the power
sorting done in step S4, not the arrangement in terms of
frequency.
[0072] After step S9, control proceeds to step S10 where, if the
total power required to transmit one bit spread across the m bins
exceeds the residual power, the program terminates, otherwise
control proceeds to step S12. At step S12, a bit is allocated to
each of the m bins.
[0073] Once the procedure of FIG. 6 terminates, the array indicates
the number of bits to transmit in each bin. Note that the procedure
can, in principle, assign "spread bits" to bins that carry one or
more entire bits by themselves.
[0074] Referring to FIG. 7, a system for a channel employing four
twisted pairs will be described. Each pair of wires is an
independent interfering channel. That is, a signal in one channel
generates a corresponding signal in the other channels. In the four
twisted pair system described here and shown in FIG. 7, a single
block of data is spread across the same number of channels, in this
case four. Four data blocks, possibly represented by, for example,
a single QAM tone, are labeled A, B, C, and D. Each block is
redundantly transmitted at four separate frequencies. However,
instead of applying the signal tone directly to the channel, the
transmitter weights the signals by respective elements of
orthogonal, 4-element Walsh codes. The receiver de-weights the
received signal at the receiving end and then despreads (sums) the
de-weighted blocks to obtain the signal. The result of this is that
the induced signal caused by interference is canceled out because
the process of weighting each original signal by one code and
de-weighting the induced signal in an adjacent channel by a
different code, orthogonal to the first, causes the sums to cancel
out. For example, signal B is weighted by a first code. The signal
induced by signal B in adjacent channels, represented in FIG. 7 by
the lower case letters, in this case "b", is also weighted by the
first code. But when the signal in an adjacent channel, say channel
3 carrying signal C, arrives at the receiver, the signal C combined
with an induced portion of B weighted by the first code is then
deweighted by a second code. Thus, when signal C is despread
(summed), the induced signal in each frequency channel cancels an
induced signal in another frequency channel.
[0075] To use the invention for eliminating interference from other
channels, it is not necessary to allocate bits to channels as
described with reference to FIGS. 5 and 6. In fact, all data could
be spread across multiple frequency channels.
[0076] Note that in a channel where signal spreading is to be used
for reducing interference, one that includes multiple interfering
sub-channels, it is desirable for communications to be set up with
cognizance of how spreading is to be done on each of the
sub-channels. That is, for example, the frequency bins in each
twisted pair in a four-twisted pair bundle should all have the same
bit assignments so that the interference will be canceled out as
described above.
[0077] As a first level concept, the present invention relates to
replicating (i.e. spreading) the spectral contents of a signal over
multiple distinct parts of the physical frequency band to
compensate for high attenuation and/or high noise in those parts of
the communication channel frequency band that would otherwise not
be usable due to noise and attenuation effects. Of course, if
signal power could be increase arbitrarily, no portion of the
spectrum would be unusable, however, practical considerations and
standards, as discussed above, dictate otherwise. The effect of the
replication is that signal-processing techniques can be used in the
Receiver to recover information sent in the impaired parts of the
frequency band, and, with the correct choice of parameters (i.e.
degree of replication), the same level of performance (specifically
bit error rate) as achieved in the unimpaired parts of the band can
be achieved in portions of the band where attenuation and/or noise
are high.
[0078] The consequence of this technique is a communication signal
where the frequency-domain spreading (replication) is used
selectively, such that not all spectral content has the same degree
of redundancy, some having no redundancy as appropriate. This might
be termed a hybrid modulation because of the mixed-mode, or
non-uniform replication. The advantage clearly lies in the fact
that spectrum is used efficiently with non-redundant transmission
for those parts of the band where adequate signal-to-noise can be
achieved, while spreading is applied to those parts of the band
where transmission would otherwise be impossible. The added
bandwidth incrementally consumed by spreading is therefore not
wasteful since it exploits portions of the band that are otherwise
unusable.
[0079] As mentioned the bit allocation process is not limited by
the aggregate power constraint that applies, but by the second
limit on the power in each band, the power spectral density mask.
This occurs may occur in long lines and/or noisy environments. The
mask prevents assigning any further bits to unoccupied bands
because the attenuation and noise would require too much power in
any of the remaining bands.
[0080] The proposed concept is to overcome the circumstance
outlined above by repeating the same X value in multiple positions
in the input array to the IFFT. In this way, signals which
individually do not have adequate signal/noise ratio to achieve the
desired bit error rate are simply summed at the receiver to achieve
the requisite SNR. As an example, summing two copies of the same
QAM vector, each contaminated with the same level of noise power
but with the respective noise uncorrelated, results in 6 dB signal
enhancement versus 3 dB. noise growth, for a net improvement of 3
dB. In general the net enhancement is increased 3 dB. each time the
number of frequency bands is doubled.
[0081] A further constraint on the allocation process described
above would be to limit the 4-tuples selected to be adjacent in
frequency, that is, the 4-tuples ( . . . in this example, m-tuples
in general) would form a sequential series of frequency bands. This
would require staring at bands with indices that are integral
multiples of 4. These constraints, while compromising ideal
allocation objectives, would tremendously increase the likelihood
that neighboring interfering channels would have 4-tuples assigned
to the same set of frequency bands. If neighboring channels are
synchronized, and have replicated vectors weighted by orthogonal,
4-element Walsh codes, the result is that the mutual interference
among this group of 4 wire pairs can be zero in any of the 4-band
spread channels. Because the services provided by such modems is
expected to be ultimately interference-limited if the service
becomes popular and widespread, this capability to avoid/mitigate
mutual interference among cooperating neighbors is significant.
Even without orthogonality, the spreading provides a significant
signal processing gain with respect to. uncorrelated interference.
The orthogonality feature adds a further enhancement.
[0082] Although according to the embodiments described above, power
is distributed equally to the bins across which a single block is
transmitted, it is clear from the description that other
possibilities exist. For example, the m bins could be assigned
power on a pro rata basis according to the noise-power/gain ratio
corresponding to each bin. Thus, the bin corresponding to the
lowest noise-power/gain ratio could carry the greatest share of the
power. Bins could also be allocated different power levels
according to the power margin available below the PSD mask
limit.
[0083] Also, although according to the embodiments described above,
a number of bins for spreading is established before allocation of
data to the bins, it is possible in view of the teachings of the
present specification, to spread different blocks of data over
different numbers of bins. For example, one block consisting of one
bit could be spread over 2 blocks and second bit, perhaps in a more
impaired (by noise and/or attenuation) portion of the spectrum
spread over a larger number of bins. This could be done iteratively
and the number of bins increased until all available spectrum were
fully exploited (within the external and practical
constraints).
[0084] Also, although according to the above embodiments, spreading
is only done when the noise-attenuation floor makes transmission of
a single bit impossible for any channels remaining, other
possibilities exist in view of the teachings of the present
specification. For example, spreading can be done in certain
frequency channels which otherwise might possibly be used to carry
a single bit on their own.
[0085] The following applications, filed concurrently herewith, are
hereby incorporated by reference:
[0086] 1. A Hybrid Fiber Twisted-pair Local Loop Network Service
Architecture (Gerszberg 41-3-13);
[0087] 2. Dynamic Bandwidth Allocation for use in the Hybrid Fiber
Twisted-pair Local Loop Network Service Architecture (Gerszberg
42-4-14);
[0088] 3. The VideoPhone (Gerszberg 43-9-2);
[0089] 4. VideoPhone Privacy Activator (Gerszberg 44-10-3);
[0090] 5. VideoPhone Form Factor (Gerszberg 45-11-4);
[0091] 6. VideoPhone Centrally Controlled User Interface With User
Selectable Options (Gerszberg 46-12-5);
[0092] 7. VideoPhone User Interface Having Multiple Menu
Hierarchies (Gerszberg 47-13-6);
[0093] 8. VideoPhone Blocker (Gerszberg 79-38-26);
[0094] 9. VideoPhone Inter-com For Extension Phones (Gerszberg
48-14-7);
[0095] 10. Advertising Screen Saver (53-17);
[0096] 11. VideoPhone FlexiView Advertising (Gerszberg
49-15-8);
[0097] 12. VideoPhone Multimedia Announcement Answering Machine
(Gerszberg 73-32-20);
[0098] 13. VideoPhone Multimedia Announcement Message Toolkit
(Gerszberg 74-33-21);
[0099] 14. VideoPhone Multimedia Video Message Reception (Gerszberg
75-34-22);
[0100] 15. VideoPhone Multimedia Interactive Corporate Menu
Answering Machine Announcement (Gerszberg 76-35-23);
[0101] 16. VideoPhone Multimedia Interactive On-Hold Information
Menus (Gerszberg 77-36-24);
[0102] 17. VideoPhone Advertisement When Calling Video Non-enabled
VideoPhone Users (Gerszberg 78-37-25);
[0103] 18. Motion Detection Advertising (Gerszberg 54-18-10);
[0104] 19. Interactive Commercials (Gerszberg 55-19);
[0105] 20. VideoPhone Electronic Catalogue Service (Gerszberg
50-16-9);
[0106] 21. A Facilities Management Platform For Hybrid Fiber
Twisted-pair Local Loop Network, Service Architecture (Barzegar
18-56-17);
[0107] 22. Multiple Service Access on Single Twisted-pair (Barzegar
(16-51-15);
[0108] 23. Life Line Support for Multiple Service Access on Single
Twisted-pair (Barzegar 17-52-16);
[0109] 24. A Network Server Platform (NSP) For a Hybrid Fiber
Twisted-pair (HFTP) Local Loop Network Service Architecture
(Gerszberg 57-4-2-2-4);
[0110] 25. A Communication Server Apparatus For Interactive
Commercial Service (Gerszberg 58-20-11);
[0111] 26. NSP Multicast, PPV Server (Gerszberg 59-21-12);
[0112] 27. NSP Internet, JAVA Server and VideoPhone Application
Server (Gerszberg 60-5-3-22-18);
[0113] 28. NSP WAN Interconnectivity Services for Corporate
Telecommuters (Gerszberg 71-9-74-21-6);
[0114] 29. NSP Telephone Directory White-Yellow Page Services
(Gerszberg 61-64-23-19);
[0115] 30. NSP Integrated Billing System For NSP services and
Telephone services (Gerszberg 62-7-5-24-20);
[0116] 31. Network Server Platform/Facility Management Platform
Caching Server (Gerszberg 63-8-6-3-5);
[0117] 32. An Integrated Services Director (ISD) For HFTP Local
Loop Network Service Architecture (Gerszberg 72-36-22-12);
[0118] 33. ISD and VideoPhone Customer Premise Network (Gerszberg
64-25-34-13-5);
[0119] 34. ISD Wireless Network (Gerszberg 65-26-35-14-6);
[0120] 35. ISD Controlled Set-Top Box (Gerszberg 66-27-15-7);
[0121] 36. Integrated Remote Control and Phone (Gerszberg
67-28-16-8);
[0122] 37. Integrated Remote Control and Phone User Interface
(Gerszberg 68-29-17-9);
[0123] 38. Integrated Remote Control and Phone Form Factor
(Gerszberg 69-30-18-10);
[0124] 39. VideoPhone Mail Machine (Attorney Docket No.
3493.73170);
[0125] 40. Restaurant Ordering Via VideoPhone (Attorney Docket No.
3493.73171);
[0126] 41. Ticket Ordering Via VideoPhone (Attorney Docket No.
3493.73712);
[0127] 42. Multi-Channel Parallel/Serial Concatenated Convolutional
Codes And Trellis Coded Modulation Encode/Decoder (Gelblum
4-3);
[0128] 43. Spread Spectrum Bit Allocation Algorithm (Shively
19-2);
[0129] 44. Digital Channelizer With Arbitrary Output Frequency
(Helms 5-3);
[0130] 45. Method And Apparatus For Allocating Data Via Discrete
Multiple Tones (filed Dec. 12, 1997, Attorney Docket No.
3493.20096--Sankaranaraya- nan 1-1);
[0131] 46. Method And Apparatus For Reducing Near-End Cross Talk In
Discrete Multi-Tone Modulators/Demodulators (filed Dec. 12, 1997,
Attorney Docket No. 3493.37219--Helms 4-32-18).
* * * * *