Method for voice recognition

• US 8,751,145 B2
• Filed: 11/30/2005
• Issued: 06/10/2014
• Est. Priority Date: 11/30/2005
• Status: Expired due to Fees

• Current Assignee: Audi AG (Volkswagen AG), Volkswagen Of America (Volkswagen AG)
• Original Assignee: Audi AG (Volkswagen AG), Volkswagen Of America (Volkswagen AG)
• Inventors: Bergmann, Carsten, Imam, M. Kashif, Gruchalski, Gerd, Prieto, Ramon Eduardo, Lathrop, William B., Mhrle, Markus
• Primary Examiner(s): Mawari, Redhwan K
• Application Number:
  US11/290,365
• Publication Number:
  US 20070124057A1
• Time in Patent Office: 3,114 Days
• Field of Search: 701/408, 701/410, 701/409, 701/412, 701/428, 701/462, 340/995.19, 340/988, 340/907, 340/995.13, 704/246, 704/270, 704/275
• US Class Current: 701/408
• CPC Class Codes: G10L 15/26 Speech to text systems G10L15/08 takes precedence
• CPC Class Codes: G10L 2015/025 Phonemes, fenemes or fenones being the recognition units
• CPC Class Codes: G10L 2015/085 Methods for reducing search complexity, pruning
• CPC Class Codes: G10L 2015/088 Word spotting

Abstract

A voice recognition method that is used for finding a street uses a database including information about a plurality of streets. The streets are characterized by respective street names and street types. A user provides a voice input for the street that the user tries to find. The voice input includes a street name and a street type. The street type is recognized by processing the voice input. Streets having the recognized street type are then selected from the database and a street name of at least one of the streets selected from the database is recognized by processing the voice input.

Claims

1. A method for voice recognition, comprising:

providing a database including information for a plurality of streets, the streets being characterized by respective street names and street types;

performing a location-based grammar reduction by eliminating streets remote from a current location of a vehicle;

providing a voice input including a street name information and a street type information;

recognizing, with a voice recognition system, a given street type selected from a given number of different street types by processing the voice input, wherein the step of recognizing a given street type is performed prior to the step of recognizing a street name;

selecting, with the voice recognition system, streets from the database that have the given street type and that have not been eliminated by the location-based grammar reduction as streets remote from the current location of the vehicle, wherein the step of selecting streets having the given street type is performed prior to the step of recognizing a street name; and

subsequently recognizing, with the voice recognition system, a street name of at least one of the streets selected from the database as having the given street type by processing the voice input.

2. The method according to claim 1, which comprises:

calculating a current location of a vehicle;

selecting streets within a given distance from the vehicle; and

recognizing a street type and a street name of only the streets within the given distance from the vehicle.

3. The method according to claim 1, which comprises:

performing a phoneme recognition on the voice input and selecting at least one phoneme recognized with a given confidence level; and

recognizing the street name by matching the at least one phoneme recognized with the given confidence level to phonemes characterizing the streets stored in the database.

4. The method according to claim 1, which comprises:

performing a phoneme recognition on the voice input and selecting at least one phoneme recognized with a given confidence level;

recognizing a number of street names by matching the at least one phoneme recognized with the given confidence level to phonemes characterizing the streets stored in the database; and

outputting the recognized streets.

5. The method according to claim 1, which comprises storing the voice input as a data file and recognizing the street type and the street name by processing the data file.

6. The method according to claim 1, which comprises recognizing the street type and the street name without requiring a voice input for information about a city.

7. The method according to claim 1, which comprises:

performing a phoneme recognition on each segment of a speech waveform, the speech waveform representing the voice input; and

selecting streets having at least one phoneme recognized with a given confidence level.

8. A method for finding information in a database, comprising:

storing street information in a database, the street information including a street name and a street type for each of a plurality of streets;

eliminating streets from the database that are remote from a given location;

recognizing, with a voice recognition system, a given street type selected from a given number of different street types by processing a voice input, wherein the step of recognizing a given street type is performed prior to the step of recognizing a street name;

eliminating streets from the database that have a street type different from the given street type; and

subsequently recognizing, with the voice recognition system, a street name by matching at least one phoneme recognized in the voice input to at least one phoneme characterizing respective streets that have not been eliminated from the database.