Method for natural voice recognition based on a generative transformation/phrase structure grammar

• US 7,548,857 B2
• Filed: 06/26/2003
• Issued: 06/16/2009
• Est. Priority Date: 06/28/2002
• Status: Expired due to Fees

• Current Assignee: T-Mobile Deutschland GmbH (Deutsche Telekom AG)
• Original Assignee: Minebea Company Limited (MinebeaMitsumi Inc.)
• Inventors: Markefka, Gunthert, Liedtke, Klaus Dieter
• Primary Examiner(s): ABEBE, DANIEL DEMELASH
• Application Number:
  US10/519,653
• Publication Number:
  US 20060161436A1
• Time in Patent Office: 2,182 Days
• Field of Search: 704/1, 704/9, 704/200, 704/231, 704/246, 704/251, 704/257, 704/275, 704/270, 704/270.1, 707/3, 707/4
• US Class Current: 704/246
• CPC Class Codes: G10L 15/1815 Semantic context, e.g. disambiguation of the recognition hypotheses based on word meaning
• CPC Class Codes: G10L 15/19 Grammatical context, e.g. disambiguation of the recognition hypotheses based on word sequence rules
• CPC Class Codes: Y10S 707/99933 Query processing, i.e. searching
• CPC Class Codes: Y10S 707/99934 Query formulation, input preparation, or translation

Abstract

The invention relates to a method for natural voice recognition based on a generative transformation/phrase structure grammar known as GT/PS grammar. According to the invention, a spoken phrase is analyzed for triphones contained therein, words contained in the spoken phrase are formed from the recognized triphones with the aid dictionaries and the spoken phrase is syntactically reconstructed from the recognized words using a grammar. The GT/PS grammar is a novel method enabling target sentences to be placed in said grammar. It uses traditional Grammar Specification Language (GSL), structures said sentences however in an innovative manner. It is oriented towards the rules of phrase structure grammar and Noam Chomsky'"'"'s concept of generative transformation grammar.

Claims

1. A method for natural voice recognition based on a generative transformation/phrase structure grammar, comprising the following steps:

analyzing a spoken phrase for triphones contained therein;

forming words, contained in the spoken phrase, from the recognized triphones with the aid of dictionaries; and

syntactically reconstructing the spoken phrase from the recognized words using a grammar, characterized in that the syntactic reconstruction of the spoken phrase comprises the following steps;

allocating the recognized words to part-of-speech categories, including verbs, nouns, etc.;

allocating the part-of-speech categories to nominal phrases and verbal phrases;

combining the nominal phrases and verbal phrases according to syntactic rules into an object having a sequence of part-of-speech categories;

comparing the sequence of the object having the sequence of part-of-speech categories with a plurality of sequences of part-of-speech categories of predetermined sentence models, and, in the case of an agreement, a sentence is considered as recognized and an action in a voice controlled application is triggered, wherein each predetermined sentence model has a number of variables allocated to part-of-speech categories, and when a sentence is considered as recognized, the variables allocated to the part-of-speech categories of the recognized sentence are filled with corresponding part-of-speech categories of the recognized words.

2. The method as claimed in claim 1, characterized in that the words to be recognized are held available subdivided into various part-of-speech categories in the dictionaries.

3. The method as claimed in claim 1, characterized in that the objects or parts thereof are linked to corresponding action parameters of a voice-controlled application.

4. The method as claimed in claim 1, wherein each predetermined sentence model is defined by a word order of the part-of-speech categories.

5. The method as claimed in claim 4, wherein the word order of the part-of-speech categories of the predetermined sentence model comprises “

subject, verb, object”

or “

object, verb, subject.”