Classifications

• • H—ELECTRICITY
  • H04—ELECTRIC COMMUNICATION TECHNIQUE
  • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
  • H04L5/00—Arrangements affording multiple use of the transmission path
  • H04L5/003—Arrangements for allocating sub-channels of the transmission path
  • H04L5/0058—Allocation criteria
  • H04L5/0064—Rate requirement of the data, e.g. scalable bandwidth, data priority
• • H—ELECTRICITY
  • H04—ELECTRIC COMMUNICATION TECHNIQUE
  • H04B—TRANSMISSION
  • H04B3/00—Line transmission systems
  • H04B3/02—Details
  • H04B3/32—Reducing cross-talk, e.g. by compensating
• • H—ELECTRICITY
  • H04—ELECTRIC COMMUNICATION TECHNIQUE
  • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
  • H04L5/00—Arrangements affording multiple use of the transmission path
  • H04L5/003—Arrangements for allocating sub-channels of the transmission path
  • H04L5/0044—Arrangements for allocating sub-channels of the transmission path allocation of payload
  • H04L5/0046—Determination of how many bits are transmitted on different sub-channels
• • H—ELECTRICITY
  • H04—ELECTRIC COMMUNICATION TECHNIQUE
  • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
  • H04L5/00—Arrangements affording multiple use of the transmission path
  • H04L5/003—Arrangements for allocating sub-channels of the transmission path
  • H04L5/0058—Allocation criteria
  • H04L5/006—Quality of the received signal, e.g. BER, SNR, water filling
• • H—ELECTRICITY
  • H04—ELECTRIC COMMUNICATION TECHNIQUE
  • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
  • H04L5/00—Arrangements affording multiple use of the transmission path
  • H04L5/0001—Arrangements for dividing the transmission path
  • H04L5/0003—Two-dimensional division
  • H04L5/0005—Time-frequency
  • H04L5/0007—Time-frequency the frequencies being orthogonal, e.g. OFDM(A), DMT

Definitions

• the invention relates to a method and apparatus for minimizing near end cross talk due to discrete multi-tone (DMT) transmission in cable binders.
• DMT discrete multi-tone
• QAM quadrature amplitude modulation
• a transmission frequency band is separated into N sub-bands or frequency bins, each corresponding to one QAM channel.
• each sub-band has a different capacity as a result of the variation of noise and attenuation with frequency.
• 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.
• the power spectral density mask may be dictated by a standard used in a particular country implementing the standard (such as A.N.S.I. standard Tl.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 a 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.
• 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, for example, in the megahertz of spectrum), the constraint may be -34 dBm/Hz.
• the attenuation/noise characteristics of a medium can be graphically represented by a floor in a power spectral graph where frequency may be represented in 100 kHz increments.
• 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 transmit 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 is 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 external communications standard.
• the other limit is on the aggregate power, also defined by an external communication standard; for example, ANSI Standard T 1.413- 1995 limits the total or aggregate power for all sub-bands to 100 m Watts in the downstream direction.
• 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 Tl.413-1995, may also require a minimum number of bits in a frequency band if the band is to convey any information at all.
• the power spectral density mask limit may require that less energy be used than the minimum required to transmit a single bit.
• 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.
• 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.
• 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 C. 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.
• the applicable metaphor could be described as an ice-cube pouring analogy.
• DSL Digital Subscriber loop
• FEXT Far End Cross Talk
• NXT Near End Cross Talk
• SNR channel signal-to-noise ratio
• the transmitter sends a known pseudo noise (PN) signal.
• PN pseudo noise
• the receiver computes the characteristics of the received signal in the form of a ratio N k /g k .
• g k is the channel gain (inverse of the attenuation) in frequency band k
• N 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.
• the optimum approach for a 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.
• the g k N 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).
• DSL modems that use the Discrete Multi-Tone (DMT) technology must use an algorithm for assigning data bits to the multiple tones that are used for modulation.
• the algorithms in the prior art do not specifically attempt to minimize the near end cross talk at the central office, although the implicit intentions of the algorithm designers include the objective of minimizing the NEXT. All other algorithm designers have assumed that since the NEXT is linearly proportional to the transmitted power, minimizing the transmit power is sufficient to minimizing the crosstalk.
• the coupling of the transmit power in a typical cable binder of pairs of twisted copper wires from a source pair to a disturber pair within the binder is a non-linear function of frequency. Therefore, equal power values in two different frequency bins result in different levels of cross-talk power.
• the invention solves this problem by altering the performance function that is used by the conventional algorithms.
• Figure 2 shows a modem in communication over one or more twisted wire pairs for purposes of describing an embodiment of the invention
• Figure 3 shows a general diagram of elements of a communication system for purposes of describing an embodiment of the invention
• Figure 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
• Figure 5 illustrates the major processing steps in the operation of a DMT modem
• Figure 6 schematically shows the various inputs to a bit allocation algorithm. Detail Description
• a transmitting modem 31 is connected to a receiving modem 32 by a cable 33 having one twisted pair of conductors.
• cable 33 having one twisted pair of conductors.
• high signal attenuation at higher frequencies greater than 500 kHz.
• This characteristic of a typical cable 33 is represented graphically by curve A' in FIG. 1A.
• source encoder 12 compresses data from data source 11 and applies the result to channel encoder 13 for error correction/detection data and applies the result to digital modulator 14.
• Digital modulator 14 acts as the interface with the communication channel 15 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 received 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.
• a parallel-to-serial converter 43 adds a cyclic prefix (one known method of preventing intersymbol interference) and passes the resulting data stream through an D/A converter 44 yielding a single analog signal. 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.
• a cyclic prefix one known method of preventing intersymbol interference
• Figure 5 shows the major processing steps in the operation of a DMT modem.
• the transmitter and the receiver modems perform a process of measurement 510 during which each of the two modems at the telecommunications central office (CO) and the remote terminal (RT) measure the power spectral density (PSD) of the noise over the bandwidth of interest.
• each modem transmits a known pseudorandom noise (PRN) sequence for a pre-determined duration while the respective receiving modem remains silent.
• PRN pseudorandom noise
• the received signal is recorded for subsequent processing.
• the second step 520 in the initialization processing consists of estimation of the channel transfer functions by each of the two modems.
• the noise PSDs and the channel transfer functions are inputs to the bit and power allocation algorithm.
• settings for any time and frequency domain equalizers can be estimated and set in the respective modems.
• Data transmission 530 commences after the initialization is completed.
• the model of the PSD of the NEXT that could be caused by n identical power sources within a binder is given by the followingexpression for binders containing up to 50 pairs of wires.
• ⁇ j. is the noise power
• H* is the channel transfer function at frequency k.
• the following two expressions represent the increment in power needed for a bit-addition algorithm and the decrement in power obtained in a bit-removal algorithm, respectively.
• the bit-removal algorithm would sort the array of ⁇ E quantities in the descending order. From now on, the N ⁇ XT- minimization approach will be described in terms of a bit-addition algorithm. The principle applies equally well to the previously discussed bit-removal algorithm.
• the bit-addition algorithm requires that the array of ⁇ E ⁇ be sorted in ascending order and a bit added to the frequency bin that requires the least increment of power.
• the power needed for increasing the number of bits by one bit in the frequency bin, to which a bit was added in this step, is recalculated, and the array is re-sorted.
• Bit allocation is then determined based on this array, such that bit allocation to frequency bins requires the least possible power for the maximum possible or the desired data rate. This process is continued until the addition of one bit in any of the frequency bins violates at least one of the constraints - power budget, power mask and maximum number of bits per frame.
• the algorithm described above would minimize the total power consumed for transmitting the maximum possible number of data bits per frame, while staying within all the required constraints.
• a modification of the incremental power term would result in an algorithm that can minimize NEXT, instead.
• the incremental power term is weighted by f k and then the array of these weighted terms is used to determine the bit allocation. Note that the constant ; used in Eq. 1 need not be applied since it will be common to all terms and hence, will not affect the sorting order.
• the effect of the weighting function would be to force the allocation towards the lower frequencies.
• the algorithm would result in a total integrated NEXT power that would be less than the integrated
• FIG. 6 schematically shows the various inputs to the bit allocation algorithm.
• the blocks 61-68 at the top of the figure represent the various objective functions that can be identified from the various algorithms that exist in the literature and include the new objective functions (represented by shaded blocks) that are defined in this document. Any algorithm that minimizes a performance function can utilize the cross-talk dependent function defined herein.
• Figure 6 also includes a performance function block 63 that is left undefined to indicate the possibility that other performance functions are also possible.
• Function 67 a shaded block, represents the minimization of the total Cross-talk (NEXT) power for a given data rate.
• Function 68 represents the joint minimization of the total power and maximization of the total transmitted data rate. All of the processes or functions 61-68 provide input to function selection box 69.
• the decision block 69 represents the selection of the function to be optimized for the bit and power allocation algorithm that a modem designer intends to use.
• Block 70 inputs the selected function to be optimized from block 69 to the bit and power allocation algorithm 73.
• Block 71 supplies the required constraints to the allocation algorithm 73.
• Block 72 supplies the parameters that were measured and estimated during the initial measurement and calculation processes 510 and 520 ( Figure 5).
• Block 73 represents the bit and power allocation algorithm that is implemented after all the above steps 61-72 as utilized are completed.

Landscapes

• Engineering & Computer Science (AREA)
• Signal Processing (AREA)
• Computer Networks & Wireless Communication (AREA)
• Quality & Reliability (AREA)
• Cable Transmission Systems, Equalization Of Radio And Reduction Of Echo (AREA)

Priority Applications (1)

Applications Claiming Priority (4)

Publications (1)

Family

ID=26878981

Family Applications (1)

Country Status (4)

Cited By (3)

Families Citing this family (37)

Citations (4)

Family Cites Families (2)

• 2001
  • 2001-01-26 US US09/769,516 patent/US6393052B2/en not_active Expired - Fee Related
  • 2001-02-14 TW TW90103274A patent/TW574810B/zh not_active IP Right Cessation
  • 2001-02-15 EP EP01910725A patent/EP1256223A1/en not_active Withdrawn
  • 2001-02-15 WO PCT/US2001/004835 patent/WO2001061956A1/en active Application Filing

Patent Citations (4)

Non-Patent Citations (2)

Also Published As

Similar Documents

Legal Events