Voice dialogue system, method, and program

• US 8,719,036 B2
• Filed: 03/15/2011
• Issued: 05/06/2014
• Est. Priority Date: 09/29/2010
• Status: Active Grant

• Current Assignee: Kabushiki Kaisha Toshiba (Toshiba Corporation), Toshiba Digital Solutions Corporation (Toshiba Corporation)
• Original Assignee: Kabushiki Kaisha Toshiba (Toshiba Corporation)
• Inventors: Ariu, Masahide
• Primary Examiner(s): Godbold, Douglas
• Assistant Examiner(s): ESTES, ERNEST LANGSTON
• Application Number:
  US13/048,193
• Publication Number:
  US 20120078634A1
• Time in Patent Office: 1,148 Days
• Field of Search: 704270-278, 704231-257
• US Class Current: 704/275
• CPC Class Codes: G10L 15/22 Procedures used during a speech recognition process, e.g. man-machine dialogue

Abstract

A voice dialogue system executing an operation command inputted by a voice dialogue user which stores a history of the number of times each operation is executed. Upon the reception or detection of a voice skip signal during voice input, the system ignores or skips the current voice input and acquires or retrieves an operation name with a high priority based on the history of the number of executions. The acquired operation name is then read aloud by a generation unit.

Claims

1. A voice dialogue system executing an operation by a voice dialogue with a user, by recognizing a user'"'"'s input voice and outputting a voice response corresponding the user'"'"'s input voice, comprising:

a history storage unit storing an operation name of an operation executed by the voice dialogue system and an operation history corresponding to a number of execution times of the executed operation;

a voice storage unit storing voice data corresponding to the operation name;

an input unit, comprising at least one of a button unit or a gaze detection unit, inputting a voice skip signal that is intended, by the user, to skip a user'"'"'s voice input;

a detection unit detecting the voice skip signal inputted by said input unit;

an acquisition unit acquiring the operation name of the operation having a high priority, based on a weighted largest number of execution times from said history storage unit, when said detection unit detects the voice skip signal; and

a generation unit reading the voice data corresponding to the acquired operation name from said voice storage unit, and generating a voice signal corresponding to the read voice data.

2. The voice dialogue system according to claim 1, wherein said input unit inputting the voice skip signal at a time when the voice dialog system outputs the voice response corresponding the user'"'"'s input voice.

3. The voice dialogue system according to claim 1, further comprising,

an output unit converting the voice signal into a voice, and outputting the voice.

4. The voice dialogue system according to claim 3, further comprising,

a cancel determination unit determining whether the voice skip signal is cancelled before output of the voice is finished;

and

an execution unit executing the operation corresponding to the operation name when the voice skip signal is not cancelled.

5. The voice dialogue system according to claim 4,

wherein said button unit generates the voice skip signal in a depressed state, and does not generate the voice skip signal in a non-depressed state.

6. The voice dialogue system according to claim 4, wherein

said cancel determination unit determines whether the voice skip signal is cancelled before the output of the voice is finished based on a continuous detection of the voice skip signal by said detection unit.

7. The voice dialogue system according to claim 6, further comprising,

a measuring unit measuring a duration of the voice skip signal from a time beginning when said detection unit starts to detect the voice skip signal,

wherein said cancel determination unit determines whether the voice skip signal is continued at a time when the output of the voice is finished based on the duration.

8. A voice dialogue system executing an operation by a voice dialogue with a user, by recognizing a user'"'"'s input voice and outputting a voice response corresponding the user'"'"'s input voice, comprising:

a history storage unit storing an operation name of an operation executed by the voice dialogue system and an operation history corresponding to a number of execution times of the executed operation;

a voice storage unit storing voice data corresponding to the operation name;

an input unit, comprising at least one of a button unit or a gaze detection unit, inputting a voice skip signal that is intended, by the user, to skip a user'"'"'s voice input;

a detection unit detecting the voice skip signal inputted by said input unit;

an acquisition unit acquiring the operation name of the operation having a high priority based on a weighted largest number of execution times from said history storage unit, when said detection unit detects the voice skip signal;

a generation unit reading the voice data corresponding to the acquired operation name from said voice storage unit, and generating a voice signal corresponding to the read voice data;

an output unit converting the voice signal into a voice, and outputting the voice;

an execution unit executing the operation corresponding to the operation name when the voice skip signal is not cancelled;

a measuring unit measuring a duration of the voice skip signal from a time beginning when said detection unit starts to detect the voice skip signal;

a cancel determination unit determining;

whether the voice skip signal is cancelled before output of the voice is finished based on a continuous detection of the voice skip signal by said detection unit;

or

whether the voice skip signal is continued at a time when the output of the voice is finished based on the duration;

wherein said acquisition unit acquires the number of execution times of the operation from said history storage unit and calculates a total number of execution times that is a sum of the number of execution times when the detection unit detects the voice skip signal, said acquisition unit further includes a number-of-times determination unit determining whether the total number of execution times is equal to or smaller than a predetermined threshold, and

wherein said execution unit executes the operation corresponding to the operation name if a determination of said number-of-times determination unit is true.

9. A voice dialogue system executing an operation by a voice dialogue with a user, by recognizing a user'"'"'s input voice and outputting a voice response corresponding the user'"'"'s input voice, comprising:

a history storage unit storing an operation name of an operation executed by the voice dialogue system and an operation history corresponding to a number of execution times of the executed operation;

a voice storage unit storing voice data corresponding to the operation name;

an input unit, comprising at least one of a button unit or a gaze detection unit, inputting a voice skip signal that is intended, by the user, to skip a user'"'"'s voice input;

a detection unit detecting the voice skip signal inputted by said input unit;

an acquisition unit acquiring the operation name of the operation having a high priority, that is determined based on a weighted largest number of execution times, from said history storage unit, when said detection unit detects the voice skip signal;

a generation unit reading the voice data corresponding to the acquired operation name from said voice storage unit, and generating a voice signal corresponding to the read voice data;

an output unit converting the voice signal into a voice, and outputting the voice;

a cancel determination unit determining whether the voice skip signal is cancelled before output of the voice is finished;

an execution unit executing the operation corresponding to the operation name when the voice skip signal is not cancelled; and

a management unit that transitions the dialogue state to a state corresponding to the voice data on a voice output from the output unit at a time when said detection unit finishes detecting the voice skip signal;

wherein

said voice storage unit stores the voice data which is made to correspond with a dialogue state.

10. A voice dialogue method for executing an operation by a voice dialogue with a user, comprising:

storing an operation name of an operation executed by the voice dialogue system and an operation history corresponding to a number of execution times of the executed operation, by a history storage unit;

storing voice data corresponding to the operation name by a voice storage unit;

detecting a voice skip signal indicating a skipping of a user'"'"'s voice input by a detection unit;

inputting a voice skip signal that is intended, by the user, to skip a user'"'"'s voice input, by an input unit, wherein the input unit comprises at least one of a button unit or gaze detection unit;

recognizing a user'"'"'s input voice, by a recognition unit, and outputting a voice response corresponding the user'"'"'s input voice, by an output unit;

acquiring the operation name of the operation having a high priority, that is determined based on a weighted largest of execution times, from the history storage unit, when the detection unit detects the voice skip signal, by an acquisition unit; and

reading the voice data corresponding to the acquired operation name from said voice storage unit, and generating a voice signal corresponding to the read voice data, by a generation unit.

11. The voice dialogue method according to claim 10, further comprising,

converting the voice signal into a voice, and outputting the voice by an output unit.

12. The voice dialogue method according to claim 11, further comprising,

determining whether the voice skip signal is cancelled before output of the voice is finished, by a cancel determination unit;

and

executing the operation corresponding to the operation name when the voice skip signal is not cancelled, by an execution unit.

13. The voice dialogue method according to claim 12,

wherein said button unit generates the voice skip signal in a depressed state, and does not generate the voice skip signal in a non-depressed state.

14. The voice dialogue method according to claim 12, wherein

said step of determining whether the voice skip signal is cancelled before the output of the voice is finished is based on a continuous detection of the voice skip signal, by the cancel determination unit.

15. The voice dialogue method according to claim 14, further comprising,

measuring a duration of the voice skip signal from a time beginning when said detection unit starts to detect the voice skip signal, by a measuring unit,

wherein the cancel determination unit determines whether the voice skip signal is continued at a time when the output of the voice is finished based on the duration.

16. A voice dialogue method for executing an operation by a voice dialogue with a user, comprising:

storing an operation name of an operation executed by the voice dialogue system and an operation history corresponding to a number of execution times of the executed operation, by a history storage unit;

storing voice data corresponding to the operation name by a voice storage unit;

detecting a voice skip signal indicating a skipping of a user'"'"'s voice input by a detection unit;

acquiring the operation name of the operation having a high priority, that is based on a weighted largest number of execution times, from the history storage unit, when the detection unit detects the voice skip signal, by an acquisition unit;

reading the voice data corresponding to the acquired operation name from said voice storage unit, and generating a voice signal corresponding to the read voice data, by a generation unit;

converting the voice signal into a voice, and outputting the voice by an output unit;

determining whether the voice skip signal is cancelled before output of the voice is finished, by a cancel determination unit, based on a continuous detection of the voice skip signal;

executing the operation corresponding to the operation name when the voice skip signal is not cancelled, by an execution unit,

measuring a duration of the voice skip signal from a time beginning when said detection unit starts to detect the voice skip signal, by a measuring unit, when it is determined, by the cancel determination unit, that the voice skip signal is continued at a time when the output of the voice is finished based on the duration;

the number of execution times of the operation from said history storage unit and calculating a total number of execution times that is a sum of the number of execution times when the detection unit detects the voice skip signal, by the acquisition unit; and

determining whether the total number of execution times is equal to or smaller than a predetermined threshold by a number-of-times determination unit, and

executing the operation corresponding to the operation name if said number-of-determination unit determines that the total number of execution times is equal to or larger than the predetermined threshold, by said execution unit.

17. A voice dialogue method for executing an operation by a voice dialogue with a user, comprising:

storing an operation name of an operation executed by the voice dialogue system and an operation history corresponding to a number of execution times of the executed operation, by a history storage unit;

storing voice data, which is made to correspond to a dialogue state, corresponding to the operation name, by a voice storage unit;

detecting a voice skip signal indicating a skipping of a user'"'"'s voice input by a detection unit;

acquiring the operation name of the operation having a high priority, that is determined based on a largest execution times from the history storage unit, when the detection unit detects the voice skip signal, by an acquisition unit;

reading the voice data corresponding to the acquired operation name from said voice storage unit, and generating a voice signal corresponding to the read voice data, by a generation unit;

converting the voice signal into a voice, and outputting the voice by an output unit;

determining whether the voice skip signal is cancelled before output of the voice is finished, by a cancel determination unit;

executing the operation corresponding to the operation name when the voice skip signal is not cancelled, by an execution unit; and

transitioning a dialogue state to a state corresponding to the voice data on a voice output from the output unit at the time when said detection unit finishes detecting the voice skip signal.

18. A non-transitory computer readable storage medium comprising computer executable instructions that in response to execution, cause a computing system to perform operations, comprising:

storing an operation name of the operation executed by a voice dialogue system and an operation history corresponding to a number of execution times of the executed operation;

storing voice data corresponding to the operation name;

detecting a voice skip signal indicating a skipping of a user'"'"'s voice input;

acquiring the operation name of the operation having a high priority defined by a weighted largest number of execution time based on the stored operation history, when the voice skip signal is detected; and

reading the voice data corresponding to the acquired operation name based on the stored voice data, and generates a voice signal corresponding to the read voice data.

19. The voice dialogue program according to claim 18, further comprising,

converting the voice signal into a voice, and outputting the voice;

determining whether the voice skip signal is cancelled before output of the voice is finished; and

executing the operation corresponding to the operation name when the voice skip signal is not cancelled.