Turn-taking model

• US 8,131,553 B2
• Filed: 09/21/2009
• Issued: 03/06/2012
• Est. Priority Date: 12/22/2004
• Status: Active Grant

• Current Assignee: Shadow Prompt Technology AG
• Original Assignee: Bruce Balentine, David Attwater
• Inventors: Attwater, David, Balentine, Bruce
• Primary Examiner(s): He, Jialong
• Application Number:
  US12/563,254
• Publication Number:
  US 20100017212A1
• Time in Patent Office: 897 Days
• Field of Search: 704/270, 704/275, 704/211, 704/215, 704/248, 704/253
• US Class Current: 704/270
• CPC Class Codes: G10L 15/08 Speech classification or search
• CPC Class Codes: G10L 15/22 Procedures used during a speech recognition process, e.g. man-machine dialogue
• CPC Class Codes: G10L 15/24 Speech recognition using non-acoustical features

Abstract

A method is claimed for managing interactive dialog between a machine and a user. In one embodiment, an interaction between the machine and the user is managed in response to a timing position of possible speech onset from the user. In another embodiment, the interaction between the machine and the user is dependent upon the timing of a recognition result, which is relative to a cessation of a verbalization of a desired sequence from the machine. In another embodiment, the interaction between the machine and the user is dependent upon a recognition result and whether the desired sequence was ceased or not ceased.

Claims

1. A method for managing a dialog between an interactive system configured to provide a prompt comprising one or more phrases and a user comprising the steps:

audibly providing to the user a first portion of said prompt by said interactive system wherein said prompt comprises one or more phrases;

detecting an audible input at the interactive system wherein said audible input potentially is speech response from the user, said audible input detected during the provision of a particular phrase of said one or more phrases, wherein said particular phrase comprises a prehold portion and a hold portion wherein said prehold portion occurs before said hold portion;

determining an audible input onset time during the particular phrase when onset of said audible input is detected by said interactive system;

terminating audibly providing to the user by said interactive system a remaining portion of the particular phrase when said audible input onset time occurs during said prehold portion of said particular phrase; and

continuing to audibly provide said remaining portion of said particular phrase to said user when said audible input onset time occurs during said hold portion of said particular phrase.

2. The method of claim 1 further comprising the step of:

running a prehold timer wherein said prehold timer is started in said interactive system when said particular phrase is audibly provided to the user and said prehold timer is stopped after a certain duration, wherein if said audible onset input is detected during said prehold timer, said interactive system terminates providing to said user said remaining portion of the particular phrase.

3. The method of claim 1 wherein said particular phrase further comprises a posthold portion, and wherein said interactive system continues to provide said prompt to said user when said audible input onset time occurs during said posthold portion of said particular phrase.

4. The method of claim 1 wherein a HoldTimeOut timer at the interactive system is started upon detecting onset of said audible input and has an expiration time;

and

said interactive system continues to audibly provide to the user said remaining portion of the particular phrase until said expiration time is reached.

5. The method of claim 4 wherein the expiration of HoldTimeOut timer varies based on when said onset of said audible input is detected by said interactive system.

6. A method for managing a dialog between an interactive system configured to provide a prompt comprising one or more phrases and a user comprising the steps:

audibly providing to the user a first portion of said prompt by said interactive system wherein said prompt comprises one or more phrases;

detecting an audible input at the interactive system wherein said audible input potentially is speech response from the user, said audible input detected during the provision of a particular phrase of said one or more phrases, wherein said particular phrase comprises a hold portion and a posthold portion, wherein said posthold portion occurs after said hold portion;

determining an audible input onset time during the particular phase when onset of said audible input is detected by said interactive system;

continuing to audibly provide a remaining portion of said particular phrase to said user when said audible input onset time occurs during said hold portion of said particular phrase; and

terminating audibly providing to the user by said interactive system the remaining portion of the particular phrase when said audible input onset time occurs during said posthold portion of said particular phrase.

7. The method of claim 6 further comprising the step of:

starting a HoldTimeOut timer at the interactive system upon detecting said audible input, wherein terminating audibly providing the remaining portion of the particular phrase occurs when said HoldTimeOut timer expires.