Granted Patent B2

US 7,817,745 · App. 11/336,666 · Granted Oct 19, 2010

Assignee: Adaptive Spectrum and Signal Alignment, Inc.

Abstract

Claims (76)

1. A method for compensating crosstalk by adaptively precoding data for transmission over a channel in a Digital Subscriber Line (DSL) system, the method comprising:

providing first encoded data of a first user as a first input to a multi-user precoder;

providing second encoded data of a second user as a second input to the multi-user precoder;

collecting over time, via a controller, channel information including channel changes and noise changes;

collecting over time, via the controller, information about users in the DSL system including addition of a new user and dropping off an old user;

adaptively precoding via joint processing on the first and the second encoded data via an adaptive precoding matrix of the multi-user precoder to generate first and second precoded data from the first and second encoded data respectively, the adaptive precoding based on the channel information and on the information about the users in the DSL system,

wherein the controller is configured to adapt the adaptive precoding matrix by matching channel changes observed over time, wherein the controller is further configured to adapt the adaptive precoding matrix to changes in a noise spatial correlation of the channel, and

wherein the controller is further configured to update the adaptive precoding matrix in response to the addition of the new user in the DSL system or in response to the dropping off of the old user from the DSL system; and

transmitting the first precoded data and the second precoded data over the channel.

2. The method of claim 1 , wherein adaptively precoding the first encoded data and the second encoded data comprises applying a matrix-Q filter to the first encoded data and the second encoded data to generate the first precoded data and the second precoded data respectively, and wherein the matrix-Q filter is updated, based on the channel information collected by the controller, to account for changes in the DSL system.

3. The method of claim 1 , wherein adaptively precoding the first encoded data and the second encoded data comprises:

applying a user-variable feedback filter to the first encoded data and the second encoded data to generate filtered data; and

applying a matrix-Q filter to the filtered data to generate the first precoded data and the second precoded data,

wherein at least one of the following is updated, based on the channel information collected by the controller, to account for changes in the DSL system:

the user-variable feedback filter, and

the matrix-Q filter.

4. The method of claim 1 , wherein adaptively precoding the first encoded data and the second encoded data comprises:

applying a user-variable feedback filter to the encoded data to generate filtered data;

performing modulo arithmetic on the filtered data to generate wrapped data; and

applying a matrix-Q filter to the wrapped data to generate the first precoded data and the second precoded data,

wherein at least one of the following is updated, based on the channel information collected by the controller, to account for changes in the DSL system:

the user-variable feedback filter,

the matrix-Q filter, and

the modulo arithmetic.

5. The method of claim 1 , wherein adaptively precoding the first encoded data and the second encoded data comprises:

applying a user-variable feedback filter to the encoded data to generate filtered data;

applying a dither signal to the filtered data to generate randomized data;

performing modulo arithmetic on the randomized data to generate wrapped data; and

applying a matrix-Q filter to the wrapped data to generate the first precoded data and the second precoded data,

wherein at least one of the following is updated, based on the channel information collected by the controller, to account for changes in the DSL system:

the user-variable feedback filter,

the matrix-Q filter,

the modulo arithmetic, and

the dither signal.

6. The method of claim 1 , wherein adaptively precoding the first encoded data and the second encoded data comprises dynamically precoding the first encoded data and the second encoded data by applying a user-variable feedback filter to the encoded data to generate the first precoded data and the second precoded data only when the first precoded data and the second precoded data would violate a power limitation.

7. The method of claim 6 , wherein the power limitation is one of:

a maximum transmit power, or

an average transmit power.

8. The method of claim 4 , wherein the modulo arithmetic is applied by reducing the transmit power of the first precoded data and the second precoded data.

9. The method of claim 1 further comprising:

ordering data on a plurality of DSL lines in the DSL system for a sequence in which the data will be precoded, wherein the ordering comprises:

precoding the data bound for receivers that do not use subtraction-only precoding or mod-enabled precoding;

precoding the data bound for receivers that use subtraction-only precoding; and

precoding the data bound for receivers that use mod-enabled precoding.

10. The method of claim 1 further comprising:

receiving at a single-user receiver one of the first precoded data and the second precoded data transmitted over the channel.

11. The method of claim 6 , wherein the power limitation is violated when a ratio of signal power to crosstalk power exceeds a power ratio threshold.

12. The method of claim 10 , wherein the single-user receiver includes a modulo arithmetic for maximizing a detection constellation d min during detection in the single-user receiver.

13. The method of claim 12 , wherein precoding the first encoded data and the second encoded data to generate the first precoded data and the second precoded data further comprises adding a dither signal prior to performing modulo arithmetic in the transmitter, and

wherein the dither signal is removed from the single-user receiver prior to performing a modulo arithmetic in the single-user receiver.

14. A Digital Subscriber Line (DSL) apparatus comprising:

a multi-user adaptive precoder to compensate crosstalk on a channel coupled with the DSL apparatus; and

a controller coupled with the multi-user adaptive precoder to collect over time channel information from the channel including channel changes and noise changes, and to collect information about users in the DSL system including addition of a new user and dropping off an old user,

wherein the controller to adaptively precode via joint processing on the first and the second encoded data via an adaptive precoding matrix of the multi-user precoder to generate first and second precoded data from the first and second encoded data respectively, the adaptive precoding based on the channel information and on the information about the users in the DSL system,

wherein the controller is configured to adapt the adaptive precoding matrix by matching channel changes observed over time

wherein the controller is further configured to adapt the adaptive precoding matrix to changes in a noise spatial correlation of the channel, and

wherein the controller is further configured to update the adaptive precoding matrix in response to the addition of the new user in the DSL system or in response to the dropping off of the old user from the DSL system.

15. The DSL apparatus of claim 14 , wherein the controller updates a multi-user matrix-Q filter of the multi-user adaptive precoder.

16. The DSL apparatus of claim 14 , wherein the multi-user adaptive precoder comprises:

a multiplication unit having an input and an output, wherein the multiplication unit to apply a multi-user matrix-Q filter to a data signal provided at the multiplication unit input to generate a precoded data signal at the multiplication unit output; and

a subtraction unit having an output coupled with the multiplication unit input, wherein the subtraction to apply a multi-user variable feedback filter to an encoded data signal at the subtraction unit input to generate a filtered data signal at the subtraction unit output.

17. The DSL apparatus of claim 16 , wherein the controller disables the subtraction unit when a power condition is violated by application of the multi-user variable feedback filter to the encoded data signal.

18. The DSL apparatus of claim 16 further comprising:

a mod operator having an input coupled with the subtraction unit, wherein the mod operator to apply modulo arithmetic to the filtered data signal to reduce transmit power and generate a wrapped data signal at an output of the mod operator.

19. The DSL apparatus of claim 18 further comprising:

a dither signal generator coupled with the subtraction unit and the mod operator, wherein the dither signal generator to insert a dither signal into the filtered data signal prior to application of modulo arithmetic by the mod operator.

20. The DSL apparatus of claim 14 , further comprising:

a transmitter to transmit the first precoded data and the second precoded data over the channel.

21. The DSL apparatus of claim 14 , wherein the adaptive precoding matrix is repeatedly adapted by the controller to match the changes in noise observed on the channel over time.

22. The DSL apparatus of claim 14 , wherein the adaptive precoding matrix is repeatedly adapted by the controller to match the changes in noise observed on the channel over time based on the updated coefficient values received by the controller.

23. The DSL apparatus of claim 14 , wherein the DSL apparatus comprises a multi-user vectored DSL apparatus.

24. The method of claim 1 , wherein the adaptive precoding matrix is repeatedly adapted by the controller to match the channel changes over time.

25. The method of claim 1 , wherein the adaptive precoding matrix is repeatedly adapted by the controller to match the changes in noise observed on the channel over time based on the updated coefficient values received by the controller.

26. The method of claim 1 , wherein the DSL system comprises a multi-user vectored DSL system.

27. The method of claim 1 , wherein the adaptations to the adaptive precoding matrix comprise employing matrix-dense vector multiplication techniques which derive computational savings through exploiting triangular matrix properties observed within the adaptive precoding matrix, the triangular matrix properties comprising the adaptive precoding matrix having some of its elements equal to zero (0) or sufficiently close to zero (0).

28. The method of claim 4 , wherein the modulo arithmetic is applied by minimizing the transmit power of the first precoded data and the second precoded data.

Assignments (10)

Continuity (3)