Monitoring and control of an adaptive filter in a communication system

• US 20040001450A1
• Filed: 06/24/2002
• Published: 01/01/2004
• Est. Priority Date: 06/24/2002
• Status: Active Grant

• Current Assignee: NXP USA, Inc. (NXP Semiconductors NV)
• Original Assignee: Freescale Semiconductor, Inc. (NXP Semiconductors NV)
• Inventors: Smith, Roger A., Dyba, Roman A., He, Perry P., Pessoa, Lucio F.C.
• Granted Patent: US 7,242,762 B2
• Time in Patent Office: Days
• US Class Current: 370/286
• CPC Class Codes: H04B 3/23 using a replica of transmitted signal in the time domain, e.g. echo cancellers

Abstract

A communication system having an echo canceller is disclosed. One embodiment of the echo canceller includes an adaptive filter used to provide an estimate of reflected echo which is removed from the send signal. The echo canceller may also include a near-end talker signal detector which may be used to prevent the adaptive filter from adapting when a near-end talker signal is present. The echo canceller may also include a nonlinear processor used to further reduce any residual echo and to preserve background noise. The echo canceller may also include a monitor and control unit which may be used to monitor the filter coefficients and gain of the adaptive filter to maintain stability of the echo canceller, estimate pure delay, detect a tone, and inject a training signal. The echo canceller may also include a nonadaptive filter used to reduce the length of the adaptive filter.

Claims

1. A method for performing near-end signal detection, comprising:

determining power of an output of a receiving path;

determining power of an error signal; and

using the power of the output of the receiving path and the power of the error signal to detect if the near-end signal is present.

2. The method as claim 1, wherein the step of using comprises:

forming a ratio of the power of the error signal to the power of the output of the receiving path; and

comparing the ratio to a first predetermined threshold.

3. The method as claim 2, wherein the step of using further comprises:

comparing the ratio to a second predetermined threshold.

4. The method as claim 3, wherein the step of using further comprises:

selecting one of the first predetermined threshold and the second predetermined threshold based upon a state of an adaptive filter.

5. The method as claim 3, wherein the step of using further comprises:

selecting one of the first predetermined threshold and the second predetermined threshold based upon a value of echo return loss enhancement (ERLE).

6. The method as claim 1, further comprising:

determining power of an input of a sending path;

determining a minimum of the power of the error signal and the power of the input of the sending path; and

using the minimum and the power of the output of the receiving path to detect if the near-end signal is present.

7. The method as claim 1, further comprising:

operating an echo canceller which uses the near-end signal detection.

8. The method as claim 7, further comprising:

using the echo canceller in a communication system having at least one echo path and at least one hybrid.

9. The method as claim 8, wherein the method for performing near-end signal detection is independent of a delay of the echo path and is independent of echo return loss of the hybrid.

10. The method as claim 7, wherein the step of operating the echo canceller comprises:

determining a state of an adaptive filter;

storing a plurality of filter coefficients when the state of the adaptive filter is a first predetermined state; and

performing no storing of the plurality of filter coefficients when the state of the adaptive filter is a second predetermined state.

11. A method as in claim 7, further comprising:

halting adaptation of at least a portion of a plurality of filter coefficients when a parameter of the near-end signal is below a first predetermined threshold and a parameter of a far-end signal is below a second predetermined threshold.

12. A method as in claim 7, wherein the echo canceller comprises an adaptive filter, the method further comprising:

monitoring a plurality of filter coefficients in the adaptive filter;

detecting a predetermined pattern in the plurality of filter coefficients; and

if the predetermined pattern is detected, loading a different plurality of filter coefficients into the adaptive filter.

13. A method as in claim 8, further comprising:

selectively adjusting a gain of the error signal; and

in response to selectively adjusting the gain of the error signal, ensuring that a gain of the communication system is not greater than one.

14. A method as in claim 1, further comprising:

filtering a receiving path signal with a first DC notch filter to produce the output of the receiving path;

filtering a sending path signal with a second DC notch filter to produce an input of the sending path; and

combining the output of the receiving path and the input of the sending path to produce the error signal.

15. A method for reducing divergence of a filter having a plurality of filter coefficients, wherein the filter has a plurality of states which comprise a first state and a second state, the method comprising:

determining a current state of the adaptive filter;

storing the plurality of filter coefficients when the current state of the adaptive filter is the first state; and

performing no storing of the plurality of filter coefficients when the current state of the adaptive filter is the second state.

16. A method as in claim 15, further comprising:

providing a state machine to model the plurality of states of the adaptive filter.

17. A method as in claim 15, further comprising:

restoring the plurality of filter coefficients to a previous set of coefficient values in response to detection of a near-end signal.

18. A method as in claim 15, further comprising:

restoring the plurality of filter coefficients to a previous set of coefficient values, wherein the previous set of coefficient values is not a most recently stored set of coefficient values.

19. A method as in claim 15, wherein only two sets of coefficient values are stored at any given time for the plurality of filter coefficients.

20. A method for reducing divergence of a filter having a plurality of filter coefficients, the method comprising:

halting adaptation of at least a portion of the plurality of filter coefficients when a parameter of a near-end signal is below a first predetermined threshold and a parameter of a far-end signal is below a second predetermined threshold.

21. A method as in claim 20, wherein the parameter of the near-end signal comprises power and the parameter of the far-end signal comprises power.

22. A method as in claim 20, wherein the filter is used in a communication system comprising at least one hybrid, the method further comprising:

using the plurality of filter coefficients to detect a change in the hybrid.

23. A method for improving stability of an adaptive filter, comprising:

monitoring a plurality of filter coefficients in the adaptive filter;

detecting a predetermined pattern in the plurality of filter coefficients; and

if the predetermined pattern is detected, loading a different plurality of filter coefficients into the adaptive filter.

24. A method as in claim 23, wherein the step of loading the different plurality of filter coefficients into the adaptive filter comprises:

resetting the adaptive filter to a predetermined state.

25. A method as in claim 24, wherein the predetermined state of the adaptive filter is an initial state, the method further comprising:

continuing adaptation of the adaptive filter.

26. A method as in claim 24, wherein the predetermined state of the adaptive filter is a bypass state, the method further comprising:

halting adaptation of the adaptive filter.

27. A method as in claim 23, wherein the adaptive filter is used in a communication system, and wherein the predetermined pattern occurs when the communication system is open-loop.

28. A method for improving stability of an adaptive filter for use in a communication system, the method comprising:

selectively adjusting a gain of an error signal; and

in response to selectively adjusting the gain of the error signal, ensuring that a gain of the communication system is not greater than 1.

29. A method as in claim 28, wherein the step of selectively adjusting the gain of the error signal comprises:

determining power of an input of a sending path;

determining power of the error signal; and

using the power of the input of the sending path and the power of the error signal to selectively adjust the gain of the error signal.

30. An echo canceller, comprising:

a near-end signal detector which determines a power of an output of a receiving path and a power of an error signal to detect if a near-end signal is present;

an adaptive filter, coupled to the near-end signal detector;

control circuitry, coupled to the adaptive filter; and

a non-linear processor, coupled to the control circuitry.

31. The echo canceller as in claim 30, wherein the control circuitry determines a current state of the adaptive filter, stores a plurality of filter coefficients of the adaptive filter when the current state of the adaptive filter is a first predetermined state, and restores the plurality of filter coefficients to a previous set of coefficients values when the current state of the adaptive filter is a second predetermined state.

32. The echo canceller as in claim 30, wherein the control circuitry monitors a plurality of filter coefficients in the adaptive filter, detects a predetermined pattern in the plurality of filter coefficients, and in response to detecting the predetermined pattern loads a different plurality of filter coefficients into the adaptive filter.

33. The echo canceller as in claim 30, further comprising:

gain control, coupled to the control circuitry, which selectively adjusts a gain of the error signal.

34. The echo canceller as in claim 30, wherein the adaptive filter has a plurality of filter coefficients, wherein the control circuitry halts adaptation of at least a portion of the plurality of filter coefficients when a parameter of the near-end signal is below a first predetermined threshold and a parameter of a far-end signal is below a second predetermined threshold.