System, device and method for time-domain equalizer training using an auto-regressive moving average model

• US 6,674,795 B1
• Filed: 04/04/2000
• Issued: 01/06/2004
• Est. Priority Date: 04/04/2000
• Status: Expired due to Fees

• Current Assignee: Avaya Incorporated
• Original Assignee: Nortel Networks Limited (Nortel Networks Corporation)
• Inventors: Purkovic, Aleksandar, Liu, Qingli
• Primary Examiner(s): Vo, Don N.
• Assistant Examiner(s): GHULAMALI, QUTBUDDIN
• Application Number:
  US09/542,212
• Time in Patent Office: 1,372 Days
• Field of Search: 375/231, 375/229-236, 708/323, 708/445
• US Class Current: 375/231
• CPC Class Codes: H03H 21/0012 Digital adaptive filters
• CPC Class Codes: H04L 2025/03414 Multicarrier
• CPC Class Codes: H04L 2025/03643 Order recursive
• CPC Class Codes: H04L 25/03038 with a non-recursive structure H04L25/03031 takes precedence

Abstract

A system, device, and method for time-domain equalizer (TEQ) training determines the TEQ order and TEQ coefficients by applying the multichannel Levinson algorithm for auto-regressive moving average (ARMA) modeling of the channel impulse response. Specifically, the TEQ is trained based upon a received training signal. The received training signal and knowledge of the transmitted training signal are used to derive an autocorrelation matrix that is used in formulating the multichannel ARMA model. The parameters of the multichannel ARMA model are estimated via a recursive procedure using the multichannel Levinson algorithm. Starting from a sufficiently high-order model with a fixed pole-zero difference, the TEQ coefficients corresponding to a low-order model are derived from those of a high-order model.

Claims

1. A channel shortening method comprising:

within a communication system, modeling a communication channel using an auto-regressive moving average model that is converted into a two-channel auto-regressive model through embedding;

solving the parameters of the auto-regressive moving average model using a multichannel Levinson algorithm; and

deriving coefficients for a low-order model from coefficients of a high-order model with a fixed pole-zero difference.

2. The channel shortening method of claim 1, wherein the high-order model is based upon a pole-zero model having p poles and q zeros, and wherein p is greater than q.

3. The channel shortening method of claim 1, wherein the high-order model is based upon a pole-zero model having p poles and q zeros, and wherein p is greater than a predetermined minimum value.

4. An apparatus comprising a time-domain equalizer for equalizing a communication channel and training logic for determining time-domain equalizer coefficients, wherein the training logic is operably coupled to model the communication channel using an auto-regressive moving average model that is converted into a two-channel auto-regressive model through embedding and solve the parameters of the auto-regressive moving average model using a multichannel Levinson algorithm, wherein coefficients are derived for a low-order model from coefficients of a high-order model with a fixed pole-zero difference.

5. A program product for determining time-domain equalizer coefficients, the program product comprising a computer readable medium having embodied therein a computer program, the computer program comprising:

logic for modeling a communication channel using an auto-regressive moving average model that is converted into a two-channel auto-regressive model through embedding;

logic for solving the parameters of the auto-regressive moving average model using a multichannel Levinson algorithm; and

logic for deriving coefficients for a low-order model from coefficients of a high-order model with a fixed pole-zero difference.

6. A communication system comprising a first communication device in communication with a second communication device over a communication channel, wherein the first communication device is operably coupled to transmit a training signal to the second communication device over the communication channel, and wherein the second communication device is operably coupled to receive the training signal over the communication channel and determine time-domain equalizer coefficients based upon the received training signal using an auto-regressive moving average model that is converted into a two-channel auto-regressive model through embedding and solving the parameters of the auto-regressive moving average model using a multichannel Levinson algorithm and wherein coefficients are derived for a low-order model from coefficients of a high-order model with a fixed pole-zero difference.

7. A method for detecting within a communication system instability in a multichannel Levinson algorithm, the multichannel Levinson algorithm comprising a plurality of recursions, each recursion producing a matrix having a non-zero entry only at the upper diagonal location, the method comprising:

monitoring within the communication system said non-zero entry;

terminating recursions if said non-zero entry becomes less than or equal to zero; and

terminating recursions if said non-zero entry becomes greater than or equal to one.