Process and device for the prediction of noise contained in a received signal
View Patent ↗A device and method for predicting noise in a signal that is received by a digital receiver is provided. The device includes an autocorrelation function determination device and a noise predictor device. Also included in the device is an adaptive filter that produces a prediction error value for the received signal and the adaptive filter adjusts the filter coefficients as a function of a prediction error of the adaptive filter. Using the filter coefficient of the adaptive filter, the method may determine the autocorrelation function of the received signal. The noise estimated values for the noise contained in the received signal may be computed on a basis of the autocorrelation function of the received signal. The embodiment also may facilitate noise prediction when heavy correlated noise is present.
1. A device for predicting noise in a received signal, comprising:
an autocorrelation function determination device that determines an autocorrelation function of the received signal;
a noise predictor device that determines noise estimated values for the noise contained in the received signal from the autocorrelation function of the received signal by computing the noise estimated values from an autocorrelation function of the noise contained in the received signal, the autocorrelation function of the noise being determined by subtracting an autocorrelation function of a received wanted signal from the autocorrelation function of the received signal, the autocorrelation function of the received wanted signal being derived from received signal power of the received wanted signal; and
an adaptive filter receiving the received signal and producing a prediction value for the received signal based on the received signal, wherein the adaptive filter has filter coefficients and adaptively adjusts the filter coefficients as a function of a prediction error of the adaptive filter, and the autocorrelation determination device derives the autocorrelation function of the received signal from the filter coefficients of the adaptive filter when the adaptive filter is in a stable state,
wherein the autocorrelation function of the received wanted signal is determined as follows:
r
v
(
τ
)
=
{
P
v
τ
=
0
0
τ
≠
0
,
where r v (τ) designates the autocorrelation function of the received wanted signal and P V , a received signal power of the received wanted signal, and
wherein the received signal is a pulse amplitude-modulated signal and 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:
P
v
=
1
3
M
2
-
1
M
2
,
where M designates the number of stages of the pulse amplitude modulation.
2. The device of claim 1 , where the autocorrelation function determination device further comprises an autocorrelation filter for directly determining the autocorrelation function of the received signal from the received signal.
3. The device of claim 1 , further comprising a digital receiver for receiving the received signal transmitted via a transmission channel.
4. A method for predicting noise in a received signal, comprising the following steps:
(a) providing the received signal to an adaptive filter;
(b) producing a prediction value for the received signal with the adaptive filter on a basis of the received signal;
(c) adjusting a filter coefficient of the adaptive filter dependent upon a prediction error of the adaptive filter;
(d) determining an autocorrelation function of the received signal based on the adjusted filter coefficient of the adaptive filter; and
(e) computing noise estimated values for the noise contained in the received signal on a basis of the autocorrelation function of the received signal, wherein the autocorrelation function of the received signal is determined from the adjusted filter coefficient of the adaptive filter, where the adaptive filter has reached a stable state, including computing the noise estimated values from an autocorrelation function of the noise contained in the received signal, wherein the autocorrelation of the noise is determined from the autocorrelation function of the received signal, the autocorrelation function of the noise being determined by subtracting an autocorrelation function of a received wanted signal, on which the received signal is based, from the autocorrelation function of the received signal, wherein the autocorrelation function of the received wanted signal is derived from the received signal power of the received wanted signal,
wherein the autocorrelation function of the received wanted signal is determined as follows:
r
v
(
τ
)
=
{
P
v
τ
=
0
0
τ
≠
0
,
where r v (τ) designates the autocorrelation function of the received wanted signal and P V , a received signal power of the received wanted signal, and
wherein the received signal is a pulse amplitude-modulated signal and 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:
P
v
=
1
3
M
2
-
1
M
2
,
where M designates the number of stages of the pulse amplitude modulation.
5. The method of claim 4 , wherein the determining step uses a value of a symbol received with the received signal using a trellis diagram.
6. The method of claim 5 , further comprising the step of correcting a metric of the trellis diagram as a function of the noise estimated values.
7. The method of claim 4 , further comprising the step of equalizing the received signal before determining the autocorrelation function of the received signal.
8. The method of claim 4 , further comprising the step of adjusting the filter coefficients of the adaptive filter as a function of the prediction error using an LMS algorithm.
9. The method of claim 4 , wherein the adaptive filter is an adaptive line enhancer filter.
10. The method of claim 4 , wherein the autocorrelation function of the received signal is determined from the filter coefficients of the adaptive filter according to the 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 in the stable state of the filter, {right arrow over (r)} a autocorrelation vector of the received signal and R an autocorrelation matrix of the received signal on the basis of the autocorrelation vector.
11. The method of claim 10 , wherein the autocorrelation function of the received signal is obtained from the equation using a reversive Levinson-Durbin algorithm.
12. The method of claim 4 , further comprising the step of determining directly from the received signal the autocorrelation function of the received signal using an autocorrelation filter.
13. The method of claim 4 , wherein the received signal is modulated and 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.
14. The method of claim 4 , wherein the autocorrelation function of the noise is determined from the following:
r
n
(
τ
)
=
{
r
u
(
0
)
-
P
v
τ
=
0
r
u
(
τ
)
τ
≠
0
,
where r n (τ) is the autocorrelation function of the noise, r u (τ) is the autocorrelation function of the received signal and P v a received signal power of the received wanted signal.
15. The method of claim 4 , further comprising the step of obtaining noise predictor coefficients for a noise predictor from an 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, R n an autocorrelation matrix of the noise on the basis of the autocorrelation vector and {right arrow over (w)} n a vector with the noise predictor coefficients.
16. The method of claim 15 , further comprising the step of determining the noise predictor coefficients via the equation using the reversive Levinson-Durbin algorithm.
17. The method of claim 15 , further comprising the step of adaptively adjusting the noise predictor coefficients after the coefficients have been determined.
18. A method for predicting noise in a received signal, comprising the following steps:
(a) providing the received signal to an adaptive filter;
(b) producing a prediction value for the received signal with the adaptive filter on a basis of the received signal;
(c) adjusting a filter coefficient of the adaptive filter dependent upon a prediction error of the adaptive filter;
(d) determining an autocorrelation function of the received signal based on the adjusted filter coefficient of the adaptive filter, where the adaptive filter has reached a stable state;
(e) processing the autocorrelation function of the received signal for considering a correlation of the received signal by a wanted signal contained in the received signal; and
(f) determining noise estimated values for the noise contained in the received signal from the processed autocorrelation function of the received signal, including determining the noise estimated values with the autocorrelation function of the received signal wherein only a section of the autocorrelation function of the received signal is evaluated, the received signal substantially containing the entirety of information about the noise,
wherein prior to determining the noise estimated values from the evaluated section of the autocorrelation function of the received signal, the autocorrelation function of the evaluated section of the autocorrelation function of the received signal is again constituted to determine the noise estimated values on the basis of this re-constituted autocorrelation function.
19. The method of claim 18 , wherein the autocorrelation function of the received signal is processed by using a Tomlinson Harashima pre-coding to which a transmitted signal based on the received signal is subjected.
20. The method of claim 18 , wherein the section of the autocorrelation function of the received signal to be evaluated is selected so that the section substantially contains the entire amplitude information and frequency information about the noise.
21. The method of claim 18 , wherein the section of the autocorrelation function r u (τ) of the received signal to be evaluated corresponds to a section of the autocorrelation function of the received signal where τ is greater than or equal to a specific limit value.
22. The method of claim 18 , wherein when only the section of the autocorrelation function of the received signal is evaluated, a phase offset in the autocorrelation function of the received signal is corrected.
23. The method of claim 18 , wherein the autocorrelation function of the evaluated section of the autocorrelation function of the received signal comprises a convolution of the evaluated section of the autocorrelation function of the received signal.
24. The method of claim 18 , wherein prior to determining the noise estimated values, the autocorrelation function of the evaluated section of the autocorrelation function of the received signal is scaled with a scaling factor.
25. A device for predicting noise in a received signal, comprising:
an autocorrelation function determination device that determines an autocorrelation function of the received signal;
a noise predictor device that processes the autocorrelation function of the received signal for considering a correlation of the received signal by a wanted signal contained in the received signal and determines noise estimated values for the noise contained in the received signal from the processed autocorrelation function of the received signal, the noise predictor device configured to determine the noise estimated values with the autocorrelation function of the received signal wherein only a section of the autocorrelation function of the received signal is evaluated, the received signal substantially containing the entirety of information about the noise, and wherein prior to determining the noise estimated values from the evaluated section of the autocorrelation function of the received signal, the autocorrelation function of the evaluated section of the autocorrelation function of the received signal is again constituted to determine the noise estimated values on the basis of this re-constituted autocorrelation function; and
an adaptive filter receiving the signal and producing a prediction value for the received signal based on the received signal, wherein the adaptive filter has filter coefficients and adaptively adjusts the filter coefficients as a function of a prediction error of the adaptive filter, and the autocorrelation determination device derives the autocorrelation function of the received signal from the filter coefficients of the adaptive filter when the adaptive filter is in a stable state.