Process and device for the prediction of noise contained in a received signal

• US 20050195925A1
• Filed: 11/19/2004
• Published: 09/08/2005
• Est. Priority Date: 11/21/2003
• Status: Active Grant

• Current Assignee: Maxlinear Incorporated
• Original Assignee: Lantiq Deutschland GMBH (Intel Corporation)
• Inventors: Traber, Mario
• Granted Patent: US 7,616,714 B2
• Time in Patent Office: Days
• US Class Current: 375/343
• CPC Class Codes: H04L 1/0054 Maximum-likelihood or sequential decoding, e.g. Viterbi, Fano, ZJ algorithms
• CPC Class Codes: H04L 1/006 Trellis-coded modulation
• CPC Class Codes: H04L 1/20 using signal quality detector
• CPC Class Codes: H04L 2025/03496 as a prediction filter

Abstract

Process and device for the prediction of noise contained in a received signal For predicting noise (n) contained in a received signal (u) transmitted via a transmission channel (2), it is proposed that the autocorrelation function of the received signal (u) is obtained in order to determine noise estimated values for the noise (n) contained in the received signal (u) on the basis of the autocorrelation function. On the basis of the noise-estimated values determined thereby metrics can be updated in a decider in the form of a Viterbi decoder (6), in order to make correct decisions.

Claims

1. Process for the prediction of noise contained in a received signal, comprising the following steps:

(a) determining the autocorrelation function of the received signal (u), and

(b) computing noise estimated values for the noise (n) contained in the received signal (u) on the basis of the autocorrelation function determined in step (a) of the received signal.

2. Process according to claim 1, characterized in that the decision is made by way of a value of a symbol received with the received signal (u) using a trellis diagram, in which metrics of the trellis diagram are corrected as a function of the noise estimated values.

3. Process according to claim 1, characterized in that the received signal (u) is equalized before the autocorrelation function has been determined.

4. Process according to claim 1, characterized in that the received signal (u) is fed to an adaptive filter (14), which produces prediction values for the received signal on the basis of the received signal (u) wherein, dependent on a prediction error of the adaptive filter (14), filter coefficients of the adaptive filter (14) are adjusted, and in that the autocorrelation function of the received signal (u) is determined from the filter coefficients of the adaptive filter (14), if the adaptive filter (14) has reached a stable state.

5. Process according to claim 4, characterized in that the filter coefficients of the adaptive filter (14) are adjusted as a function of the prediction error (e) using an LMS algorithm.

6. Process according to claim 4, characterized in that the adaptive filter (14) is an “

adaptive line enhancer”

filter.

7. Process according to claim 4, characterized in that the autocorrelation function of the received signal (u) is determined from the filter coefficients of the adaptive filter (14) according to the following equation:

R{right arrow over (w)}₀={right arrow over (r)},

where {right arrow over (w)}₀ designates a vector with the filter coefficients of the adaptive filter (14) in a stable state of the filter (14), {right arrow over (r)} the autocorrelation vector of the received signal (u) and R the autocorrelation matrix of the received signal (u) on the basis of the autocorrelation vector.

8. Process according to claim 7, characterized in that the autocorrelation function of the received signal (u) is obtained from the equation indicated using the reversive Levinson-Durbin algorithm.

9. Process according to claim 1, characterized in that the autocorrelation function of the received signal (u) is directly determined from the received signal using an autocorrelation filter (12).

10. Process according to claim 1, characterized in that the autocorrelation function of the noise (n) contained in the received signal (u) is determined from the auto correlation function of the received signal (u), in order to determine the noise estimated values from the autocorrelation function of the noise (n).

11. Process according to claim 10, characterized in that the autocorrelation function of the noise (n) is determined by subtraction of the autocorrelation function of a received wanted signal, on which the received signal (u) is based, from the autocorrelation function of the received signal (u).

12. Process according to claim 11, characterized in that the autocorrelation function of the received wanted signal is determined as follows:

13. Process according to claim 12, characterized in that the received signal (u) is pulse amplitude-modulated, wherein to determine the autocorrelation function of the received wanted signal the received signal power P_v of the received wanted signal is determined dependent on a number of stages of the pulse amplitude modulation as follows:

14. Process according to claim 12, characterized in that the received signal is modulated, and in that to determine the autocorrelation function of the received wanted signal the received signal power of the received wanted signal is determined dependent on the method of modulation.

15. Process according to claim 11, characterized in that the autocorrelation function of the noise (n) is determined as follows:

16. Process according to claim 10, characterized in that noise predictor coefficients for a noise predictor are obtained from the autocorrelation function of the noise for determining the noise-estimated values via the following equation:

R_n{right arrow over (w)}_n={right arrow over (r)}_n,

where {right arrow over (r)}_n designates the autocorrelation vector of the noise (n), R_n the autocorrelation matrix of the noise (n) on the basis of the autocorrelation vector and {right arrow over (w)}_n a vector with the noise predictor coefficients.

17. Process according to claim 16, characterized in that the noise predictor coefficients are determined via the equation indicated using the reversive Levinson-Durbin algorithm.

18. Process according to claim 16, characterized in that after the noise predictor coefficients have been determined these are subsequently adaptively adjusted.

19. Process according to claim 1, characterized in that the autocorrelation function of the received signal (u) is processed before the noise estimated values are obtained from the autocorrelation function of the received signal (u) for considering a correlation of the received signal (u) by a wanted signal contained in the received signal (u), wherein the noise estimated values are determined from the processed autocorrelation function of the received signal (u).

20. Process according to claim 19, characterized in that the autocorrelation function of the received signal (u) is processed for considering a Tomlinson Harashima pre-coding, to which a transmitted signal (v) based on the received signal (u) has been subjected.

21. Process according to claim 19, characterized in that the autocorrelation function of the received signal (u) is processed in such a way that to determine the noise estimated values only a section of the autocorrelation function of the received signal (u) is evaluated, which received signal contains substantially entirely noise information about the noise (n).

22. Process according to claim 21, characterized in that the section of the autocorrelation function of the received signal (u) to be evaluated is selected in such a way that it contains substantially entire amplitude information and frequency information of the noise (n).

23. Process according to claim 21, characterized in that the section of the autocorrelation function r_u(τ

) of the received signal (u) to be evaluated corresponds to a section of the autocorrelation function of the received signal (u), where τ

≧

a specific limit value.

24. Process according to claim 21, characterized in that when the autocorrelation function of the received signal (u) is evaluated a phase offset in the autocorrelation function of the received signal (u) is corrected in consequence of the evaluation of only the section of the autocorrelation function of the received signal (u).

25. Process according to claim 21, characterized in that before the noise estimated values have been determined from the evaluated section of the autocorrelation function of the received signal (u) the autocorrelation function of the evaluated section of the autocorrelation function of the received signal (u) is again constituted, in order to determine the noise estimated values on the basis of this re-constituted autocorrelation function.

26. Process according to claim 25, characterized in that the autocorrelation function of the evaluated section of the autocorrelation function of the received signal (u) is constituted in that convolution of the evaluated section of the autocorrelation function of the received signal (u) is carried out.

27. Process according to claim 25, characterized in that before the noise estimated values have been determined the autocorrelation function of the evaluated section of the autocorrelation function of the received signal (u) is scaled with a scaling factor.

31. Device according to claim 28, characterized in that the device is configured for executing the process according to any one of claims 1-27.

28. Device for the prediction of noise contained in a received signal,

with an autocorrelation function determination device (12) for determining the autocorrelation function of the received signal (u), and

with a noise predictor device (6, 12) for determining noise estimated values for the noise (n) contained in the received signal (u) from the autocorrelation function of the received signal (u).

29. Device according to claim 28, characterized in that the device comprises an adaptive filter (14), to which the received signal (u) is fed, wherein the adaptive filter (14) is configured for producing prediction values for the received signal on the basis of the received signal (u), and wherein the adaptive filter (14) is configured for the adaptive adjustment of its filter coefficients as a function of a prediction error of the adaptive filter (14), and in that the autocorrelation determination device (12) is configured in such a way that it derives the autocorrelation function of the received signal (u) from the filter coefficients of the adaptive filter (14), provided the adaptive filter (14) is in a stable state.

30. Device according to claim 28, characterized in that the autocorrelation determination device (12) comprises an autocorrelation filter for directly determining the autocorrelation function of the received signal (u) from the received signal (u).

32. Digital receiver for receiving a received signal (u) transmitted via a transmission channel (2) with a device according to claim 28 for predicting noise (n) contained in the received signal (u).