IP Library Granted Patent US 8,195,456
Granted Patent B2
US 8,195,456 · App. 12/631,699 · Granted Jun 5, 2012

Robust speech recognition based on spelling with phonetic letter families

View Patent ↗
Loading inventors, assignments & file history…
Monitor This Case
Get email alerts when status or documents change.
Order Certified Copies
Most orders are placed with the USPTO same day — all within 24 business hours.
Order via The Patent Place →
Pre-filled with this patent's details
Quick Facts
Patent No.
US 8,195,456
App. No.
12/631,699
Granted
Jun 5, 2012
Kind
B2
Abstract

A system and method for entering a destination into a navigation system, usually a vehicle navigation system, that uses phonetic letter families, or groups of letters which sound similar, to improve the reliability and accuracy of speech recognition. The method involves grouping each letter of the English alphabet into a family of letters which sound similar, such as A, J, and K. When a destination name is spelled by a user, each letter is recognized in terms of the phonetic letter family to which it belongs. This phonetic equivalent spelling is compared to the navigation database of street, city, and state names, which has also been converted to its phonetic equivalent spelling. If a match is found, the user is asked to confirm that this is the desired destination.

Claims (85)

1. A method for identifying a name stored in a computer system, said method comprising:

defining a plurality of phonetic letter families representing the letters of an alphabet, where each phonetic letter family includes one or more letters which sound similar when spoken;

providing a database of names which includes both an actual spelling of each name and a phonetic equivalent spelling of each name, where the phonetic equivalent spelling of each name is spelled using the phonetic letter family for each letter in the name;

spelling a name that a user wants to identify in the database of names;

converting each letter in the name spelled by the user to its phonetic letter family, to create a series of phonetic letters, by a speech recognition processor device;

comparing the series of phonetic letters to the phonetic equivalent spelling of each name in the database, to create a list of possible matches, by a controller device;

narrowing the list of possible matches as each additional letter is added to the series of phonetic letters, by the controller device; and

presenting the list of possible matches to the user for confirmation when the user completes the spelling of the name.

2. The method according to claim 1 wherein the alphabet is the Roman alphabet.

3. The method according to claim 2 wherein the phonetic letter families include:

a group consisting of the letters A, J, and K;

a group consisting of the letters B, C, D, E, G, P, T, V, and Z;

a group consisting of the letters I and Y;

a group consisting of the letter L;

a group consisting of the letters H, M, and N;

a group consisting of the letter O;

a group consisting of the letters Q and U;

a group consisting of the letter R;

a group consisting of the letters F, S, and X; and

a group consisting of the letter W.

4. The method according to claim 1 wherein:

the computer system is a vehicle navigation system;

the database of names is a navigation database of street addresses, cities, and states; and

the name spelled by the user is a destination address containing a street number, a street name, a city name, and a state name.

5. The method according to claim 4 wherein the destination address is resolved by:

comparing the state name as spelled by the user against a list of states in the United States to determine a matched state name;

using the matched state name to create a subset of possible city names from the navigation database, where the subset of possible city names includes all of the cities in the state designated by the matched state name;

comparing the city name as spelled by the user against the subset of possible city names to determine a matched city name;

using the matched city name to create a subset of possible street names from the navigation database, where the subset of possible street names includes all of the streets in the city designated by the matched city name;

comparing the street name as spelled by the user against the subset of possible street names to determine a matched street name;

reconstructing the destination address by appending the matched street name to the street number as spoken by the user, adding the matched city name, and adding the matched state name; and

presenting the destination address as reconstructed to the user for confirmation.

6. The method according to claim 1 wherein:

the computer system is a vehicle navigation and travel assistance system;

the database of names is a database of street addresses with city and state, including the name of the residence or business at each address; and

the name spelled by the user is a destination name comprising a name of a residence or business.

7. A system for identifying a name stored in a computer, said system comprising:

a data set containing a plurality of phonetic letter families representing the letters of an alphabet, where each phonetic letter family includes one or more letters which sound similar when spoken;

a database of names which includes both an actual spelling of each name and a phonetic equivalent spelling of each name, where the phonetic equivalent spelling of each name is spelled using the phonetic letter family for each letter in the name;

a speech recognition processor that accepts a spoken input from a user, said spoken input comprising a series of letters which spell a name which the user wants to find in the database of names;

means for converting each letter of the spoken input to its phonetic letter family to create a series of input phonetic letters;

means for storing the series of input phonetic letters;

means for comparing the series of input phonetic letters to the phonetic equivalent spelling of all names in the database to create a list of possible matches, and narrowing the list of possible matches as each additional letter is added to the series of input phonetic letters; and

an output device for presenting the list of possible matches to the user for confirmation when the user completes the spoken input.

8. The system according to claim 7 wherein the alphabet is the Roman alphabet.

9. The system according to claim 8 wherein the phonetic letter families include:

a group consisting of the letters A, J, and K;

a group consisting of the letters B, C, D, E, G, P, T, V, and Z;

a group consisting of the letters I and Y;

a group consisting of the letter L;

a group consisting of the letters H, M, and N;

a group consisting of the letter O;

a group consisting of the letters Q and U;

a group consisting of the letter R;

a group consisting of the letters F, S, and X; and

a group consisting of the letter W.

10. The system according to claim 7 wherein the computer is a vehicle navigation system.

11. The system according to claim 10 wherein the database of names is a navigation database of street addresses, cities, and states and the spoken input is a destination address containing a street number, a street name, a city name, and a state name.

12. The system according to claim 11 further comprising:

means for comparing the state name as spelled by the user against a list of states in the United States to determine a matched state name;

means for using the matched state name to create a subset of possible city names from the navigation database, where the subset of possible city names includes all of the cities in the state designated by the matched state name;

means for comparing the city name as spelled by the user against the subset of possible city names to determine a matched city name;

means for using the matched city name to create a subset of possible street names from the navigation database, where the subset of possible street names includes all of the streets in the city designated by the matched city name;

means for comparing the street name as spelled by the user against the subset of possible street names to determine a matched street name;

means for reconstructing the destination address by appending the matched street name to the street number as spoken by the user, adding the matched city name, and adding the matched state name; and

means for using the output device to present the destination address as reconstructed to the user for confirmation.

13. The system according to claim 7 wherein the computer is a vehicle navigation and travel assistance system, the database of names is a database of street addresses with city and state, including the name of the residence or business at each address, and the spoken input is a destination name comprising a name of a residence or business.

14. A vehicle navigation system comprising:

a data set containing a plurality of phonetic letter families representing the letters of an alphabet, where each phonetic letter family includes one or more letters which sound similar when spoken;

a navigation database of street addresses, street names, city names, and state names, which includes both an actual spelling of each name and a phonetic equivalent spelling of each name, where the phonetic equivalent spelling of each name is spelled using the phonetic letter family for each letter in the name;

a speech recognition processor that accept a spoken input from a user, said spoken input comprising a series of letters which spell a name which the user wants to find in the navigation database, and for converting each letter of the spoken input to its phonetic letter family to create a series of input phonetic letters;

a controller for storing the series of input phonetic letters, and for comparing the series of input phonetic letters to the phonetic equivalent spelling of all names in the database to create a list of possible matches, and narrowing the list of possible matches as each additional letter is added to the series of input phonetic letters; and

a display screen for presenting the list of possible matches to the user for confirmation when the user completes the spoken input, and for graphically presenting navigation data to the user.

15. The vehicle navigation system according to claim 14 wherein the alphabet is the Roman alphabet.

16. The vehicle navigation system according to claim 14 wherein the phonetic letter families include:

a group consisting of the letters A, J, and K;

a group consisting of the letters B, C, D, E, G, P, T, V, and Z;

a group consisting of the letters I and Y;

a group consisting of the letter L;

a group consisting of the letters H, M, and N;

a group consisting of the letter O;

a group consisting of the letters Q and U;

a group consisting of the letter R;

a group consisting of the letters F, S, and X; and

a group consisting of the letter W.

Assignments (8)
RELEASE OF SECURITY INTEREST Recorded Nov 7, 2014
From: WILMINGTON TRUST COMPANY
To: GM GLOBAL TECHNOLOGY OPERATIONS LLC
Reel/Frame 034287/0001 →
CHANGE OF NAME Recorded Feb 10, 2011
From: GM GLOBAL TECHNOLOGY OPERATIONS, INC.
To: GM GLOBAL TECHNOLOGY OPERATIONS LLC
Reel/Frame 025781/0299 →
SECURITY AGREEMENT Recorded Nov 8, 2010
From: GM GLOBAL TECHNOLOGY OPERATIONS, INC.
To: WILMINGTON TRUST COMPANY
Reel/Frame 025327/0156 →
RELEASE OF SECURITY INTEREST Recorded Nov 5, 2010
From: UAW RETIREE MEDICAL BENEFITS TRUST
To: GM GLOBAL TECHNOLOGY OPERATIONS, INC.
Reel/Frame 025315/0136 →
RELEASE OF SECURITY INTEREST Recorded Nov 4, 2010
From: UNITED STATES DEPARTMENT OF THE TREASURY
To: GM GLOBAL TECHNOLOGY OPERATIONS, INC.
Reel/Frame 025246/0234 →
SECURITY AGREEMENT Recorded Feb 25, 2010
From: GM GLOBAL TECHNOLOGY OPERATIONS, INC.
To: UNITED STATES DEPARTMENT OF THE TREASURY
Reel/Frame 023989/0155 →
SECURITY AGREEMENT Recorded Feb 25, 2010
From: GM GLOBAL TECHNOLOGY OPERATIONS, INC.
To: UAW RETIREE MEDICAL BENEFITS TRUST
Reel/Frame 023990/0001 →
ASSIGNMENT OF ASSIGNOR'S INTEREST Recorded Dec 7, 2009
From: ROESSER, ROBERT P.; HERMIZ, SARMAD Y.
To: GM GLOBAL TECHNOLOGY OPERATIONS, INC.
Reel/Frame 023614/0198 →