Voice recognition system for mobile unit

• US 20050004798A1
• Filed: 05/06/2004
• Published: 01/06/2005
• Est. Priority Date: 05/08/2003
• Status: Abandoned Application

• Current Assignee: Nissan Motor Co., Ltd.
• Original Assignee: Nissan Motor Co., Ltd.
• Inventors: Lee, Akinobu, Kaminuma, Atsunobu
• Application Number:
  US10/839,747
• Publication Number:
  US 20050004798A1
• Time in Patent Office: Days
• US Class Current: 704/250
• CPC Class Codes: G10L 15/183 using context dependencies, e.g. language models
• CPC Class Codes: G10L 15/197 Probabilistic grammars, e.g. word n-grams

Abstract

An aspect of the present invention provides a voice recognition system includes that a memory unit configured to store a statistical language dictionary which statistically registers connections among words, a voice recognition unit configured to recognize an input voice based on the statistical language dictionary, a prediction unit configured to predict, according to the recognition result provided by the voice recognition unit, connected words possibly voiced after the input voice, and a probability changing unit configured to change the probabilities of connected words in the statistical language dictionary according to the prediction result provided by the prediction unit, wherein the voice recognition unit recognizes next input voice based on the statistical language dictionary changed by the probability changing unit and wherein the memory unit, the voice recognition unit, the prediction unit and the probability changing unit are configured to be installed in the mobile unit.

Claims

1. A voice recognition system for a mobile unit comprising:

a memory unit configured to store a statistical language dictionary which statistically registers connections among words;

a voice recognition unit configured to recognize an input voice based on the statistical language dictionary;

a prediction unit configured to predict, according to the recognition result provided by the voice recognition unit, connected words possibly voiced after the input voice; and

a probability changing unit configured to change the probabilities of connected words in the statistical language dictionary according to the prediction result provided by the prediction unit,

wherein the voice recognition unit recognizes next input voice based on the statistical language dictionary changed by the probability changing unit and wherein the memory unit, the voice recognition unit, the prediction unit and the probability changing unit are configured to be installed in the mobile unit.

2. The voice recognition system of claim 1, further comprising:

a voice receiver configured to receive a voice,

wherein the memory unit storing phoneme and word dictionaries to be employed for recognizing the received voice and the statistical language dictionary which statistically registers grammar of connected words; and

the probability changing unit changes the probabilities of relationships among connected words in the statistical language dictionary according to the connected words predicted by the prediction unit.

3. The voice recognition system of claim 2, wherein:

the memory unit stores the statistical language dictionary and a plurality of network grammar language dictionaries each having a network structure to describe connections among words, word groups, and morphemes; and

the probability changing unit selects at least one of the plurality of network grammar language dictionaries appropriate for the connected words predicted by the prediction unit and increases, in the statistical language dictionary, the transition probabilities of connected words in the selected network grammar language dictionary.

4. The voice recognition system of claim 2, wherein:

the memory unit stores the statistical language dictionary and at least one network grammar language dictionary, and

the probability changing unit selects at least a node of the network grammar language dictionary appropriate for the connected words predicted by the prediction unit and increases, in the statistical language dictionary, the transition probabilities of connected words in the selected node.

5. The voice recognition system of claim 3, wherein:

the network grammar language dictionary has a tree structure involving a plurality of hierarchical levels and a plurality of nodes.

6. The voice recognition system of claim 3, wherein:

the network grammar language dictionary includes information on connections between ones selected from the group consisting of word groups, words, and morphemes and at least one selected from the group consisting of a word group, a word, and a morpheme connectable to the selected ones.

7. The voice recognition system of claim 5, wherein:

the network grammar language dictionary stores place names in a hierarchical structure starting from a wide area of places to narrow areas of places; and

the prediction unit predicts, according to the hierarchical structure, connected words representative of place names possibly voiced next.

8. The voice recognition system of claim 2, further comprising:

an information controller configured to receive the recognition result from the voice recognition unit and output information to be provided for a user, and

an information providing unit configured to provide the information output from the information controller,

wherein the prediction unit predicts, according to the information provided by the information providing unit, connected words possibly voiced next; and

the probability changing unit changes, according to the connected words predicted by the prediction unit, the probabilities of connected words in the statistical language dictionary and increases, according to the information provided by the information providing unit, the probabilities of connected words in the statistical language dictionary.

9. The voice recognition system of claim 8, wherein:

if the information output from the information controller and provided by the information providing unit has a hierarchical structure, the prediction unit predicts that words in each layer of the hierarchical structure and morphemes connectable to the words form connected words possibly voiced next; and

the probability changing unit increases the probabilities of the predicted words and morphemes in the statistical language dictionary.

10. The voice recognition system of claim 8, wherein:

if the information output from the information controller and provided by the information providing unit is a group of words or a sentence of words, the prediction unit predicts that the words in the group or sentence of words form connected words possibly voiced next; and

the probability changing unit increases the connection probabilities of the same words and morphemes in the statistical language dictionary as those contained in the group or sentence of words.

11. The voice recognition system of claim 10, wherein:

the memory unit stores a thesaurus; and

the prediction unit includes, according to the thesaurus, synonyms of the predicted words in the connected words possibly voiced next.

12. The voice recognition system of claim 10, wherein:

the voice recognition unit recognizes an input voice based on the words included in the connected words possibly voiced next are limited to subjects and predicates.

13. The voice recognition system of claim 2, further comprising:

an information controller configured to receive the recognition result from the voice recognition unit and output information to be provided for a user, and

an information providing unit configured to provide the information output from the information controller,

the prediction unit predicting, according to a history of information pieces provided by the information providing unit, connected words possibly voiced next by the user,

the probability changing unit changing, according to the connected words predicted by the prediction unit, the probabilities of connected words in the statistical language dictionary and increasing, according to the information provided by the information providing unit, the probabilities of connected words in the statistical language dictionary.

14. The voice recognition system of claim 13, wherein:

the probability changing unit changes the changed probabilities of connected words toward initial probabilities as time passes.

15. The voice recognition system of claim 14, wherein:

if a word in the connected words predicted by the prediction unit is absent in the statistical language dictionary, the probability changing unit adds the word and the probabilities of the connected words to the statistical language dictionary.