Granted Patent B2

US 7,974,181 · App. 12/138,731 · Granted Jul 5, 2011

Abstract

Claims (37)

1. A method for reducing peak-to-average power ratio (PAR) in a data signal comprising:

constructing, by a transmitter, a time-domain compensation signal from a plurality of vectors generated from PAR reduction tones (PRT), wherein each PRT generates two of the plurality of vectors, wherein constructing a time-domain compensation signal comprises constructing a desired target signal, wherein subtracting the desired target signal from the data signal results in a signal having a lower PAR than the data signal; and

applying the time-domain compensation signal to the data signal after the data signal has been digitally filtered and re-sampled, wherein each vector in the plurality of vectors is generated from one of two out-of-phase symbols associated with a PRT, wherein the one of two out-of-phase symbols associated with the PRT is processed into a time-domain signal and digitally filtered and re-sampled to generate a corresponding vector.

2. The method of claim 1 wherein the two out-of-phase symbols comprise a sine symbol and a cosine symbol associated with the PRT.

3. The method of claim 1 , wherein the constructing a time-domain compensation signal further comprises multiplying a projection matrix, P, with the desired target data signal to produce the time-domain compensation signal, wherein the projection matrix is generated from the plurality of vectors.

4. The method of claim 3 wherein P=X(X T X) −1 X T , where X is a span matrix having the plurality of vectors as its columns.

5. The method of claim 1 , wherein the constructing a time-domain compensation signal further comprises:

multiplying a pseudoinverse of a span matrix with the desired target signal to generate a coefficient vector having a plurality of components with a pair of components corresponding to a PRT, wherein the span matrix has the plurality of vectors as its columns;

scaling any components in the coefficient vector which cause a PRT to exceed a power spectral density requirement to generate a revised coefficient vector; and

multiplying the span matrix with the revised coefficient vector to generate the time-domain compensation signal.

6. The method of claim 1 , wherein the constructing a desired target signal comprises:

clipping all samples of the data signal that exceed a predetermined threshold to a predetermined target; and

setting all other samples to a zero value.

7. The method of claim 1 , wherein the constructing a desired target signal comprises:

setting all samples that do not exceed a predetermined threshold to a zero value to produce an intermediate desired target signal; and

scaling the intermediate desired target signal by a scale factor to produce the desired target signal, wherein the scale factor is selected so that subtracting the desired target signal from the data signal reduces the PAR.

8. The method of claim 7 , wherein the scale factor is selected to cause the desired target signal to meet power spectral density requirements.

9. An xDSL transmitter comprising:

an inverse fast Fourier transform (IFFT);

a cyclic prefix module;

a re-sampler;

a digital filter; and

a peak-to-average power ratio (PAR) reduction module, wherein a symbol is converted into a data signal by being processed successively by the IFFT, the cyclic prefix module, the re-sampler and the digital filter and wherein the PAR reduction module predetermines a plurality of vectors generated from PAR reduction tones (PRT) by converting two out-of-phase symbols associated with each PRT into two separate vectors with the IFFT, the cyclic prefix module, the re-sampler and the digital filter, and wherein the PAR reduction module reduces the PAR in the data signal by constructing a time-domain compensation signal from the plurality of vectors and applying the time-domain compensation signal to the data signal.

10. The system of claim 9 , wherein the two out-of-phase symbols comprise a sine symbol and a cosine symbol associated with the PRT.

11. The system of claim 9 , wherein the PAR reduction module constructs the time-domain compensation signal by constructing a desired target signal wherein subtracting the desired target signal from the data signal results in a signal having a lower PAR than the data signal.

12. The system of claim 11 , wherein the PAR reduction module constructs the time-domain compensation signal by further multiplying a projection matrix with the desired target signal to produce the time-domain compensation signal, wherein the projection matrix is generated from the plurality of vectors.

13. The system of claim 11 , wherein the PAR reduction module constructs the time-domain compensation signal by further multiplying a pseudoinverse of a span matrix with the desired target signal to generate a coefficient vector having a plurality of components each corresponding to a PRT, wherein the span matrix has the plurality of vectors as its columns;

scaling any components in the coefficient vector which cause a PRT to exceed a power spectral density requirement to generate a revised coefficient vector; and

multiplying the span matrix with the revised coefficient to generate the time-domain compensation signal.

14. The system of claim 11 , wherein the PAR reduction module constructs the desired target signal by clipping all samples of the data signal that exceed a predetermined threshold to a predetermined target; and

setting all other samples to a zero value.

15. The system claim 11 , wherein the PAR reduction module constructs the desired target signal by setting all samples that do not exceed a predetermined threshold to a zero value to produce an intermediate desired target signal; and

scaling the intermediate desired target signal to produce the desired target signal by a scale factor, wherein the scale factor is selected so that subtracting the desired target signal from the data signal reduces the PAR.

16. A system for reducing peak-to-average power ratio (PAR) in a digital subscriber line system comprising:

means for constructing a time-domain compensation signal from a plurality of vectors generated from PAR reduction tones (PRT), wherein each PRT generates two of the plurality of vectors, wherein the means for constructing a time-domain compensation signal comprises means for constructing a desired target signal wherein subtracting the desired target signal from the data signal results in a signal having a lower PAR than the data signal; and

means for applying the time-domain compensation signal to a data signal after the data signal has been digitally filtered and re-sampled, wherein each vector in the plurality of vectors is generated from one of two out-of-phase symbols associated with a PRT, wherein the one of two out-of-phase symbols associated with the PRT is processed into a time-domain signal and digitally filtered and re-sampled to generate a corresponding vector.

17. The system of claim 16 , wherein the two out-of-phase symbols comprise a sine symbol and a cosine symbol associated with the PRT.

Assignments (8)