Ingress noise reduction in a digital receiver

• US 7,826,569 B2
• Filed: 04/21/2008
• Issued: 11/02/2010
• Est. Priority Date: 04/12/2001
• Status: Expired due to Fees

• Current Assignee: Juniper Networks Incorporated
• Original Assignee: Juniper Networks Incorporated
• Inventors: Buda, Fabien, Sari, Hikmet, Popper, Ambroise
• Primary Examiner(s): Ahn; Sam K
• Application Number:
  US12/106,708
• Publication Number:
  US 20090003372A1
• Time in Patent Office: 925 Days
• Field of Search: 375/219, 375/220, 375/221, 375/223, 375/229, 375/230, 375/231, 375/232, 375/233, 375/234, 375/235, 375/236, 375/346, 375/347, 375/348, 379/406.05, 708/530, 708/534, 370/286, 370/288, 370/289, 370/290
• US Class Current: 375/346
• CPC Class Codes: G11B 20/22 for reducing distortions
• CPC Class Codes: H04B 1/123 using adaptive balancing or compensation means adaptive filter circuits and algorithms H03H
• CPC Class Codes: H04N 7/102 Circuits therefor, e.g. noise reducers, equalisers, amplifiers H04N7/108 takes precedence

Abstract

The invention provides a device for reducing ingress noise in a digital signal, comprising a noise predictor for predicting an amount of ingress noise in the digital signal based on past samples of the ingress noise, and a subtractor for subtracting the predicted amount of ingress noise from the digital signal. Channel distortion is compensated for by a noise-independent equalizer, such as a ZF equalizer, placed upstream of the noise predictor. The device may be incorporated, for example, in a cable modem termination system (CMTS) of an hybrid fiber/coax (HFC) network.

Claims

1. A device, comprising:

a noise predictor circuit to determine an amount of predicted noise (û

_n) at a time cycle n in a sample (s_n) of digital data using predictor coefficients, where the sample (s_n) comprises useful data (x_n) and an actual noise sample (u_n);

an adaptation circuit to determine the predictor coefficients of the noise predictor circuit for a future time cycle n+1 based on an adaptation step size, the predicted noise (û

_n) at the time cycle n, actual noise samples at time cycles n, n−

1, n−

2, . . . n−

M, and predictor coefficients from time cycle n, where the adaptation circuit is further to determine the predictor coefficients of the noise predictor circuit based on a multiplication of the adaptation step size with a value associated with a difference between the actual noise sample (u_n) at time sample n and the predicted noise (û

_n);

a termination system for a plurality of transmitters in a time division multiple access (TDMA) network, where the device reduces the noise in digital data received from at least one of the transmitters; and

a control unit to;

reserve time slots in which no transmitter of the plurality of transmitters is allowed to transmit towards the termination system, and

control the adaptation circuit to adapt the predictor coefficients of the noise predictor circuit during the reserved time slots.

2. The device of claim 1, further comprising:

a first subtractor to subtract the amount of predicted noise (û

_n) from the sample (s_n) of the digital data to produce subtracted data (s_n−

û

_n);

a decision circuit to compare the subtracted data (s_n−

û

_n) with a set of thresholds to produce a decided symbol (d_n) representative of the useful data (x_n) in the sample (s_n); and

a second subtractor to subtract the decided symbol (d_n) from the sample (s_n) to produce the actual noise sample (u_n).

3. The device of claim 2, further comprising:

an equalizer having noise-independent coefficients, the equalizer to compensate for inter-symbol interference in the digital data.

4. The device of claim 3, where an output of the equalizer is coupled to the first subtractor.

5. The device of claim 3, where the equalizer comprises a zero-forcing equalizer.

6. The device of claim 3, further comprising:

another adaptation circuit to adapt coefficients of the equalizer.

7. The device of claim 6, where the other adaptation circuit is to adapt coefficients of the equalizer according to a least-means-square (LMS) algorithm.

8. The device of claim 2, further comprising:

a storage unit to receive and store past actual noise samples (u<sub>n−

i</sub>) from the second subtractor, where i designates past time cycles.

9. The device of claim 1, where the adaptation circuit is to determine the predictor coefficients (c_kⁿ⁺¹) of the noise predictor circuit for a future time cycle n+1 according to:

10. The device of claim 9, where the noise predictor determines the amount of predicted noise û

_n according to;

11. The device of claim 1, where the adaptation circuit is further to determine the predictor coefficients of the noise predictor circuit for the future time cycle n+1 using a tap leakage technique.

12. The device of claim 11, where the adaptation circuit is to determine the predictor coefficients (c_kⁿ⁺¹) of the noise predictor circuit for the future time cycle n+1 according to:

13. The device of claim 1, wherein the adaptation circuit is further to minimize the mean square error (|û

_n−

u_n|²) of the actual noise sample (u_n).

14. A method performed by a noise reduction device, the method comprising:

receiving, by the noise reduction device, a sample (s_n) of digital data, where the sample (s_n) comprises useful data (x_n) and an actual noise sample (u_n);

determining, by the noise reduction device, an amount of predicted noise (û

_n) at a time cycle n in the sample (s_n) using predictor coefficients;

determining, by the noise reduction device, the predictor coefficients for a future time cycle n+1 based on an adaptation step size, the predicted noise (û

_n) at a time cycle n, actual noise samples at time cycles n, n−

1, n−

2, . . . n−

M, and predictor coefficients from time cycle n, where determining the predictor coefficients is further based on a multiplication of the adaptation step size with a value associated with a difference between the actual noise sample (u_n) at time sample n and the predicted noise (û

_n);

generating, by the noise reduction device, a decided symbol (d_n) representative of the useful data (x_n) in the sample (s_n) based on the predicted noise (û

_n);

subtracting, by the noise reduction device, the amount of predicted noise (û

_n) from the sample (s_n) of the digital data to produce subtracted data (s_n−

û

_n);

comparing, by the noise reduction device, the subtracted data (s_n−

û

_n) with a set of thresholds to generate the decided symbol (d_n) representative of the useful data (x_n) in the sample (s_n); and

subtracting, by the noise reduction device, the decided symbol (d_n) from the sample (s_n) to produce the actual noise sample (u_n).

15. The method of claim 14, where the predictor coefficients (c_kⁿ⁺¹) are determined for the future time cycle n+1 according to:

16. The method of claim 15, where the amount of predicted noise û

_n is determined according to;

17. The method of claim 14, where the predictor coefficients for the future time cycle n+1 are further determined using a tap leakage technique.

18. The method of claim 17, where the predictor coefficients (c_kⁿ⁺¹) for the future time cycle n+1 are determined according to:

19. The method of claim 14, where the sample (s_n) of digital data is received in a termination system from a plurality of transmitters in a time division multiple access (TDMA) network, the method further comprising:

reserving time slots in which no transmitter of the plurality of transmitters is allowed to transmit towards the termination system; and

determining the predictor coefficients during the reserved time slots.