Method and apparatus for answering machine detection in automatic dialing

• US 6,850,602 B1
• Filed: 03/27/2002
• Issued: 02/01/2005
• Est. Priority Date: 03/27/2002
• Status: Expired due to Term

• Current Assignee: Avaya Incorporated
• Original Assignee: Avaya Technology Corporation Miami Lakes FLA US
• Inventors: Chou, Wu
• Primary Examiner(s): WEAVER, SCOTT LOUIS
• Application Number:
  US10/107,963
• Time in Patent Office: 1,042 Days
• Field of Search: 379/67.1, 379/70, 379/79, 379/80, 379/88.01, 379/88.07, 379/62.1
• US Class Current: 379/80
• CPC Class Codes: H04M 2203/2027 Live party detection
• CPC Class Codes: H04M 3/5158 in combination with automated outdialling systems devices for signalling identity of wanted subscriber H04M1/26

Abstract

In a communications network, systems and methods using a feature other than a tone for detecting if the destination is a machine are described. An exemplary method includes the following steps: establishing a communication with a destination, detecting a response, if a response is not detected within a pre-determined time interval, determining that the destination is a machine. If a response is detected, the method further includes the following steps: determining whether the response includes a speech signal that includes a predefined term (e.g., “please leave a message”), and if it does, determining that the destination is a machine. If the response does not include a predefined term, the method may send out a query to the destination and if no response to the query is received or the response is not a speech signal, the method determines that the destination is a machine.

Claims

1. A method for use in a communications network, the method comprising the steps of:

providing a communication to a destination via a communication server;

detecting in a response detector whether there is a first response from the destination; and

determining in a classifier if the destination is a machine based on whether the first response is detected and whether at least one feature other than a tone is present in the first response;

wherein the at least one feature comprises at least one predefined term of speech.

2. The method of claim 1, wherein if the first response is not detected, the determining step indicates that the destination is a machine.

3. The method of claim 1, further comprising the step of sending a query to the destination after the providing step.

4. The method of claim 3, wherein the feature is a speech signal.

5. The method of claim 4, wherein if no speech is present, the determining step indicates that the destination is a machine.

6. The method of claim 5, wherein if a speech signal is present, the determining step indicates that the destination is not a machine.

7. The method of claim 1, wherein the feature is a speech signal that comprises the predefined term.

8. The method of claim 7, wherein if the feature is present, the determining step indicates that the destination is a machine.

9. The method of claim 1, wherein the feature is a speech signal in the first response.

10. The method of claim 1, wherein the determining step comprises searching for the tone and if the tone is not present, searching for other features.

11. A method for use in a communications network, the method comprising the steps of:

providing a communication to a destination via a communication server;

detecting in a response detector whether there is a first response from the destination; and

determining in a classifier if the destination is a machine based on whether the first response is detected and whether at least one feature other than a tone is present in the first response;

wherein the feature is a speech signal that comprises a predefined term selected from the group consisting of a word, a phrase, and a sentence.

12. A method for use in a communications network, the method comprising the steps of:

providing a communication to a destination;

detecting whether there is a first response from the destination; and

determining if the destination is a machine based on whether the first response is detected and whether at least one feature other than a tone is present in the first response;

wherein the feature is a speech signal in the first response and the determining step further comprises the step of analyzing the feature using a natural language routing technique and deciding the destination as one selected from the group of a machine, a live person, and an unknown, wherein the natural language routing technique is trained with a set of predefined possible responses from the destination.

13. A method for use in a communications network, the method comprising the steps of:

providing a communication to a destination via a communication server;

detecting in a response detector whether there is a first response from the destination; and

determining in a classifier if the destination is a machine based on whether the first response is detected and whether at least one feature other than a tone is present in the first response;

wherein the determining step further comprises the steps of;

if the feature is not present, sending a query to the destination; and

detecting whether there is a second response from the destination; and

if the second response is detected, searching for a speech signal in the second response.

14. A method for use in a communications network, the method comprising the steps of:

providing a communication to a destination;

detecting whether there is a first response from the destination; and

determining if the destination is a machine based on whether the first response is detected and whether at least one feature other than a tone is present in the first response;

wherein the determining step further comprises the steps of;

if the feature is not present, sending a query to the destination;

detecting whether there is a second response from the destination;

if the second response is detected searching for a speech signal in the second response; and

if the second response is not detected, determining that the destination is a machine.

15. A method for use in a communications network, the method comprising the steps of:

providing a communication to a destination;

detecting whether there is a first response from the destination; and

determining if the destination is a machine based on whether the first response is detected and whether at least one feature other than a tone is present in the first response;

wherein the determining step further comprises the steps of;

if the feature is not present, sending a query to the destination;

detecting whether there is a second response from the destination;

if the second response is detected, searching for a speech signal in the second response; and

if the speech signal is not present in the second response, determining that the destination is a machine.

16. A method for use in a communications network, the method comprising the steps of:

providing a communication to a destination via a communication server;

detecting in a response detector whether there is a first response from the destination; and

determining in a classifier if the destination is a machine based on whether the first response is detected and whether at least one feature other than a tone is present in the first response;

wherein the determining step comprises searching for the tone and if the tone is not present, the determining step comprises searching for a predefined term.

17. A method for use in a communications network, the method comprising the steps of:

providing a communication to a destination;

detecting whether there is a first response from the destination; and

determining if the destination is a machine based on whether the first response is detected and whether at least one feature other than a tone is present in the first response;

wherein the determining step comprises searching for the tone and if the tone is not present, the determining step comprises searching for a predefined term; and

wherein if no predefined term is present, the determining step comprises analyzing the speech signal using a natural language routing technique to produce a result which is selected from the group of machine, a live person, and an unknown.

18. The method of claim 17, wherein if the result is an unknown, the determining step sends a query to the destination, detects a second response, and determines if the second response comprises a speech signal, and if the second response comprises a speech signal, determines that the destination is not a machine.

19. An apparatus for use in a communications network, the apparatus comprising:

a communication server for providing a communication to a destination using an address of the destination;

a response detector for detecting whether there is a response from the destination; and

a classifier for determining whether the destination is a machine based on whether the response is detected and whether at least one feature in the first response is present, wherein the at least one feature comprises at least one predefined term of speech.

20. The apparatus of claim 19, wherein if the response is not detected, the classifier determines that the destination is a machine.

21. The apparatus of claim 19, further comprising a dialog engine for sending a query to the destination.

22. The apparatus of claim 21, wherein the query dialog sends the query before the detector is activated to detect the first response.

23. The apparatus of claim 22, wherein the feature is a speech signal and if the speech signal is not present the classifier determines that the destination is a machine.

24. The apparatus of claim 19 further comprising:

a service database for storing the address of the destination; and

an automatic dialer for retrieving the address of the destination from the service database.

25. A method for use in a communications network, the method comprising the steps of:

(a) providing a communication to a destination via a communication server;

(b) listening for a first response within a first pre-determined interval and ignoring the first response if detected;

(c) sending a query in the form of speech;

(d) detecting in a response detector whether there is a second response within a second pre-determined interval; and

(e) if no second response is detected, determining in a classifier that the destination is a machine.

26. The method of claim 25, wherein at the end of step (e), repeat from step (c).

27. The method of claim 25, further comprising the steps of

(f) if the second response is detected, determining whether the second response includes comprises a speech signal;

and

(g) if the second response does not comprise a speech signal, determining that the destination is a machine.

28. The method of claim 27, further comprising the step of

(h) if the second response comprises a speech signal, analyzing the content of the speech signal in the second response to determine whether the destination has an intent to continue the dialog.

29. A method for use in a communications network, the method comprising the steps of:

(a) providing a communication to a destination via a communication server;

(b) detecting in a response detector for a first response within a first pre-determined interval;

(c) if the first response comprises a tone feature, determining in a classifier that the destination is a machine; and

(d) if the first response comprises a speech signal having a predefined term determining in the classifier the type of destination based on the predefined term.

30. The method of claim 29, wherein at step (d) the destination is determined a machine.

31. A method for use in a communications network, the method comprising the steps of:

(a) providing a communication to a destination via a communication server;

(b) detecting in a response detector for a first response within a first pre-determined interval;

(c) if the first response comprises a tone feature, determining in a classifier that the destination is a machine;

(d) if the first response comprises a speech signal, analyzing the content of the speech signal; and

(e) determining in the classifier the type of destination based on the result of the analysis.

32. The method of claim 31, wherein at step (d) the method of analysis is a natural language routing technique.

33. The method of claim 31, wherein at step (e) the result is a selected one of a machine, a live person, and an unknown.