Man-machine system equipped with speech recognition device

• US 6,385,582 B1
• Filed: 05/02/2000
• Issued: 05/07/2002
• Est. Priority Date: 05/03/1999
• Status: Expired due to Fees

• Current Assignee: Pioneer Corporation
• Original Assignee: Pioneer Corporation
• Inventors: Iwata, Takahiro
• Primary Examiner(s): Dorvil, Richemond
• Assistant Examiner(s): Nolan, Daniel A.
• Application Number:
  US09/562,755
• Time in Patent Office: 735 Days
• Field of Search: 704/231-258, 704/270, 704/270.1, 704/275
• US Class Current: 704/270
• CPC Class Codes: G10L 15/22 Procedures used during a speech recognition process, e.g. man-machine dialogue

Abstract

A man-machine system equipped with a voice recognition apparatus for input words is provided. The system controls each of functions, handling each control word composed of a control command word and an object word. If only the control command word is inputted into the system as one of the input words, the system complements the corresponding object word by using an object word of the other input words recognized previously. The system further designates similar candidates and stores them in a memory. If correction is requested, the current candidate is corrected with another one of the candidates stored in the memory. The system further stores the input words in the memory, so that re-recognition according to the input words previously inputted and stored in the memory is also executed when one of the candidates is corrected.

Claims

1. A speech recognition apparatus comprising:

circuits for receiving an input word inputted according to a request from the voice recognition apparatus;

a comparator for comparing the input word with reference vocabularies related to the request, and for selecting at least one of the reference vocabularies;

a first memory for storing the selected at least one of the reference vocabularies;

a controller for executing a control word having a structure composed by a control command word and an object word controlled by the control command word;

wherein, if the input word includes the control command word only, the controller substitutes the selected at least one of the reference vocabularies already stored in the first memory for the object word and creates the control word.

2. The speech recognition apparatus according to claim 1, wherein the first memory stores a plurality of selected reference vocabularies, the controller selects a candidate from the plurality of selected reference vocabularies, and the control command word includes a command to change the candidate to another one of the plurality of selected reference vocabularies stored in the first memory.

3. The speech recognition apparatus according to claim 1, wherein if the control word is inputted into the voice recognition apparatus in a lump, the controller executes the control word.

4. The speech recognition apparatus according to claim 1, wherein the control command word includes a command to correct the object word.

5. The speech recognition apparatus according to claim 3, wherein the first memory stores a plurality of selected reference vocabularies, the controller selects a candidate from the plurality of selected reference vocabularies, and the control command word includes a command to change the candidate to another one of the plurality of selected reference vocabularies stored in the first memory.

6. The speech recognition apparatus according to claim 3, wherein the control command word includes a command to correct the object word.

7. The speech recognition apparatus according to claim 4, wherein the comparator comprises:

an analyzer for analyzing the input word;

a second memory for storing an analyzed result of the analyzer; and

circuits for evaluating a similarity between the analyzed results and the reference vocabularies related to the request.

8. The speech recognition apparatus according to claim 6, wherein the comparator comprises:

an analyzer for analyzing the input word;

a second memory for storing an analyzed result of the analyzer; and

circuits for evaluating a similarity between the analyzed results and the reference vocabularies related to the request.

9. A speech recognition apparatus comprising:

circuits for receiving an input word inputted according to a request from the voice recognition apparatus;

a comparator for comparing the input word with a plurality of reference vocabularies related to the request, and for selecting selected reference-vocabularies from said plurality of reference vocabularies, based on a similarity with the input word;

a first memory for storing the selected reference vocabularies;

a controller for selecting a candidate from the selected reference vocabularies and executing a control word having a structure composed by a control command word and an object word controlled by the control command word;

wherein the control word includes a control command word to change the candidate to another one of said selected reference vocabularies, so that a desired candidate is obtained by repeating the control command word without the object word.

10. The speech recognition apparatus according to claim 9, wherein the controller substitutes a selected reference vocabulary from the selected reference vocabularies already stored in the first memory for the object word.

11. The speech recognition apparatus according to claim 9, wherein the voice recognition apparatus is included in a navigation apparatus.

12. A speech recognition apparatus comprising:

first circuits for receiving a first input word inputted according to a first request from the voice recognition apparatus, and for receiving a second input word inputted according to a second request from the voice recognition apparatus;

an analyzer for analyzing the first input word and the second input word, and for generating a first analyzed result and a second analyzed result;

a first memory for storing the first analyzed result and the second analyzed result;

second circuits for evaluating a similarity between the first analyzed result and first reference vocabularies related to the first request, and for evaluating a similarity between the second analyzed result and second reference vocabularies related to the second request;

a comparator for selecting first selected reference vocabularies from the first reference vocabularies related to the first request, and for selecting second selected reference vocabularies from the second reference vocabularies related to the second request;

a controller for selecting a first candidate from the first selected reference vocabularies and a second candidate from the second selected reference vocabularies;

wherein the first and second candidates are selected based on first and second similarity, respectively; and

wherein when the controller changes the first candidate to a corrected first candidate, the second circuits evaluate a similarity between the second analyzed result and corrected second reference vocabularies belonging to the corrected first candidate, again after the first candidate is corrected, to select corrected second selected reference vocabularies, and selects a corrected second candidate from the corrected second selected reference vocabularies.

13. The speech recognition apparatus according to claim 12, wherein the first candidate is excluded from the first selected reference vocabularies in the case that the controller corrects the first candidate.

14. The speech recognition apparatus according to claim 12, wherein the voice recognition apparatus is included in a navigation apparatus.

15. The speech recognition apparatus according to claim 12, further comprising:

a second memory for storing the first selected reference vocabularies having the similarity to the first analyzed result;

wherein the controller corrects the first candidate by selecting another one of the first selected reference vocabularies stored in the second memory.

16. The speech recognition apparatus according to claim 12, further comprising:

a second memory for storing the second selected reference vocabularies having the similarity to the second analyzed result;

wherein the controller corrects the second candidate by selecting another one of the second selected reference vocabularies stored in the second memory.

17. The speech recognition apparatus according to claim 12, further comprising:

a second memory for storing the corrected second selected reference vocabularies having the similarity to the second analyzed result;

wherein the controller changes the corrected second candidate to another one of the corrected second selected reference vocabularies stored in the second memory.

18. The speech recognition apparatus according to claim 12, further comprising:

a third input word inputted according to a third request from the voice recognition apparatus;

wherein the voice recognition apparatus analyzes the third input word, generates a third analyzed result, evaluates a similarity between the third analyzed result and third selected reference vocabularies related to the third request, and selects one of the third selected reference vocabularies as the corrected first candidate.

19. The speech recognition apparatus according to claim 15, wherein the first candidate has the highest similarity to the first input word, and said another one of the first selected reference vocabularies has the next highest similarity to the first input word.

20. The voice recognition apparatus according to claim 16, wherein the second candidate has the highest similarity to the second input word, and the another one of the second selected reference vocabularies has the next highest similarity to the second input word.

21. The speech recognition apparatus according to claim 17, wherein the corrected second candidate has the highest similarity to the second input word, and said another one of the corrected second selected reference vocabularies has the next highest similarity to the second input word.

22. The speech recognition apparatus according to claim 18, further comprising:

a second memory for storing the third selected reference vocabularies having the similarity to the third analyzed result;

wherein the controller changes the corrected first candidate to another one of the third selected reference vocabularies stored in the second memory.

23. The speech recognition apparatus according to claim 22, wherein the corrected first candidate has the highest similarity to the third input word, and said another one of the third selected words has the next highest similarity to the third input word.