Error detection/correction and fault detection/recovery – Pulse or data error handling – Digital data error correction
Reexamination Certificate
2000-10-12
2004-05-04
Chung, Phung M. (Department: 2133)
Error detection/correction and fault detection/recovery
Pulse or data error handling
Digital data error correction
C714S712000, C375S222000
Reexamination Certificate
active
06732323
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates to multi-channel data communications, including xDSL communications. More particularly, it relates to selecting or modifing forward error correction parameters used during a communications session.
2. Description of Related Art
The emerging availability of xDSL service across existing twisted pair copper wires has been a boon for small business and residential customers. This boon has been facilitated by promulgation of xDSL standards by the International Telecommunication Union (ITU), Telecommunication Standardization Sector of the ITU (ITU-T). The family of DSL Recommendations includes the following: G.992.1, G.992.2, G.991.1, G.996.1, G.994.1, G.997.1 and G.995.1. Recommendation G.995.1 provides an overview of these standards. Recommendations G.991.1, G.992.1, G.992.2 have developed techniques for transmitting a range of bit rates over the existing copper local network from relatively short distances at high bit rates, and to long distances at relatively lower bit rates. Recommendations G.994.1, G.996.1 and G.997.1 support G.992.1 and G.992.2 by providing common handshake, management and testing procedures. These standards allow substantial flexibility in implementation. For instance, an ADSL Transceiver Unit (ATU) may be configured to support synchronous (STM) or asynchronous (ATM) data.
An important aspect of setting up an xDSL communication session is selection of Reed Solomon encoding parameters for forward error correction (FEC). Following handshakes, initialization begins. This includes channel analysis and FEC parameter exchange. While the ITU standards proscribe protocols for initialization, they do not describe how ATU equipment at the central office (ATU-C) or the remote user location (ATU-R) should act on data produced by channel analysis or how these ATUs should select FEC parameters to exchange. They set a bit error rate standard (BER) of &egr;=10
−7
, but give no guidance on selection of FEC parameters to most effectively achieve that BER. This is a weakness of the standards.
Selection of FEC parameters is important because these parameters control both forward error correction and data transmission. Reed-Solomon forward error correction encoding is computationally complex, especially when high levels of error correction flexibility are adopted. It involves the overhead of redundant data, which reduces effective throughput.
Therefore, it is desirable to develop a method for consistently selecting favorable FEC parameters, based on available channel analysis data, both for initialization and revision of FEC parameters. Parameters should be selected to satisfy a BER criteria, whether the ITU criteria or a more stringent criteria.
SUMMARY OF INVENTION
The present invention includes a method for selection of forward error correction parameters used in data communications, including computing an average signal to noise ratio of a set of communications subchannels, determining a number of effective communications subchannels loaded with data bits, and selecting one or more sets of forward error correction parameters based on the average signal to noise ratio and the number of effective communications subchannels. One aspect of the present invention is that it applies to xDSL communications using a discrete multi-tonal carrier to provide subchannels. This DMT carrier may have 96 or fewer subchannels, which may be discrete or overlapping frequency bands. The DMT carrier may be generated by application of an inverse discrete Fourier transform. The forward error correction encoding for which parameters are selected may be Reed-Solomon encoding. The parameters may be a number (S) of dmt frames in an FEC frame, and a number (R/S) of redundant FEC symbols per dmt frame. The method may further include computing predetermined net coding gains, for dimensions of the average signal to noise ratio, the number of effective communications subchannels, and for the FEC parameters S, R/S. The predetermined net coding gains may be expressed in lookup tables, bicubic splines, piecewise linear approximations or any other convenient storage format. This method may further include computing predetermined sets of preferred FEC parameters S, R/S, based on the predetermined net coding gains for pairs of average signal to noise ratio and number of effective communications subchannels. Alternatively, net coding gains may be calculated and sets of preferred FEC parameters selected after initialization begins.
The present invention may be practiced as a method for selecting forward error correction parameters to be used in multi channel communications. A method embodying aspects of the present invention may include acquiring signal to noise ratio data for a set of communications channels, identifying a subset of those channels to carry data or otherwise counting or estimating the number of such channels, computing an average signal to noise ratio for the subset of channels and selecting set(s) of FEC parameters, using the average SNR and number of data carrying channels. Aspects of this embodiment include use of Reed-Solomon encoding for FEC and selecting the parameters S and R/S. Symbol sets, represented by S may consist of dmt symbols or of data carried by multiple channels in one transmission interval. The universe of parameters S and R/S may be set by international standard and may have certain values set forth below. The computations for selecting parameters may include adjusting the average signal to noise ration by a safety margin. The selected sets of parameters may be chosen from sets identified by a sending terminal and may be forwarded from a receiving terminal to a sending terminal.
Aspects of the present invention also may be embodied in a method of selecting FEC parameters including the steps of acquiring signal to noise ratio data, estimating, identifying or counting channels to be used, computing an average SNR and selecting set(s) of FEC parameters based on SNR data and the number of channels to be used, wherein a downstream terminal selects the parameters for downstream transmission and an upstream terminal selects the parameters for upstream transmission. As in the prior method, aspects of this embodiment include use of Reed-Solomon encoding, subject to parameters S and R/S, where symbol sets may be dmt symbols or other sets of data transmitted in intervals. Acceptable FEC parameters may be set by international standard and may be constrained in one direction or the other by the sending terminal.
Another embodiment practicing aspects of the present invention may be a method of selecting FEC parameters for multi channel communications including the steps of predicting net SNR gains for pairs of FEC parameters S and R/S across ranges of SNR and number of communications channels used, storing preferred pairs of FEC parameters for at least a portion of those ranges, corresponding to predicted net SNR gains, and selecting set(s) of preferred FEC parameters. Aspects of this embedment include storing the set(s) of parameters in a table organized by SNR and number of communications channels used, wherein the SNR range includes 10-50 dB and the number of channels used includes 1-96 channels (as for G.lite downstream), 1-224 channels (as for G.dmt downsteam) or 1-26 channels (as for G.lite or G.dmt upstream.)
A further embodiment practicing aspects of the present invention may include the steps of storing functions predicting a net SNR gain for pairs of FEC parameters S and R/S across ranges of SNR and number of communications channels used, measuring an average SNR ratio for multiple channels and estimating, identifying or counting a number of channels to be used, predicting the net SNR gain for pairs of FEC parameters using the SNR and number of communication channels data, and selecting pair(s) of FEC parameters. Aspects of this embodiment may include storing the functions as bicubic splines or as a series of interpolation tables.
REFERENCES:
patent: 5479447 (1995-12-01), Chow et al.
Mitlin Vlad
Murphy Tim
Williams Richard G. C.
3Com Corporation
Befel, Jr. Ernest J.
Chung Phung M.
Dooley Matthew C.
Haynes Beffel & Wolfeld LLP
LandOfFree
Method of selecting initialization parameters for... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method of selecting initialization parameters for..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method of selecting initialization parameters for... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3196598