System and method for voice recognition

• US 8,510,103 B2
• Filed: 10/15/2010
• Issued: 08/13/2013
• Est. Priority Date: 10/15/2009
• Status: Active Grant

• Current Assignee: Paul Angott
• Original Assignee: Paul Angott
• Inventors: Angott, Paul
• Primary Examiner(s): Hudspeth, David R
• Assistant Examiner(s): Nguyen, Timothy
• Application Number:
  US12/905,703
• Publication Number:
  US 20110093261A1
• Time in Patent Office: 1,033 Days
• Field of Search: 704/205
• US Class Current: 704/205
• CPC Class Codes: G10L 15/10 using distance or distortion measures between unknown speech and reference templates

Abstract

Systems and methods are operable to associate each of a plurality of stored audio patterns with at least one of a plurality of digital tokens, identify a user based on user identification input, access a plurality of stored audio patterns associated with a user based on the user identification input, receive from a user at least one audio input from a custom language made up of custom language elements wherein the elements include at least one monosyllabic representation of a number, letter or word, select one of the plurality of stored audio patterns associated with the identified user, in the case that the audio input received from the identified user corresponds with one of the plurality of stored audio patterns, determine the digital token associated with the selected one of the plurality of stored audio patterns, and generate the output signal for use in a device based on the determined digital token.

Claims

1. A computer-implemented method for generating an output signal based on spoken user input, comprising:

storing, in a memory;

a plurality of audio patterns generated by the user, each audio pattern including a single monosyllabic representation of a command, number, letter or word having two or more syllables in a custom language wherein at least one audio pattern represents both an opening command and a closing command; and

a plurality of digital tokens each associated with at least one stored audio pattern;

receiving, from a user, at least one audio input from the custom language;

selecting, using a processor, one of the plurality of stored audio patterns associated with the user, in the case that the audio input received from the user corresponds with one of the plurality of stored audio patterns;

determining, using the processor, the digital token associated with the selected one of the plurality of stored audio patterns; and

generating, using the processor, an output signal configured to communicate the command, number, letter or word based on the determined digital token.

2. The method of claim 1, wherein each of the plurality of stored audio patterns is associated with one distinct monosyllabic audio input.

3. The method of claim 1, further comprises:

identifying a user based on user identification input; and

accessing the plurality of stored audio patterns associated with a user based on the user identification input.

4. The method of claim 3, wherein identifying the user further comprises:

identifying the user based on monosyllabic audio input.

5. The method of claim 1, further comprising:

collecting the plurality of stored audio patterns from the user by;

prompting the user to utter a monosyllabic audio specimen;

receiving the uttered monosyllabic audio specimen;

converting the uttered monosyllabic audio specimen into a digital pattern; and

storing the digital pattern of the uttered monosyllabic audio specimen as the stored audio pattern.

6. The method of claim 3, wherein the stored audio patterns and digital tokens are stored in a table and one of the plurality of stored audio patterns in the table is selected by:

comparing a digital audio input signal with each stored audio pattern associated with the identified user; and

selecting one of the plurality of stored audio patterns with the highest correspondence to the digital audio input.

7. The method of claim 6, wherein comparing the digital audio input signal with each stored audio pattern associated with the identified user is performed by calculating a correlation between the digital audio input signal and the stored audio pattern.

8. The method of claim 1, wherein the custom language includes a language distinction element operable to generate an output signal for the purpose of distinguishing discrete components of communication.

9. The method of claim 5, wherein the process of collecting the plurality of stored audio patterns requires only one collection sequence per user in order to generate an output signal based on spoken user input.

10. The method of claim 9, wherein prompting the user to utter a monosyllabic audio specimen comprises sending data representing a monosyllabic custom language element from an output control circuit to at least one of an audio and visual output port.

11. The method of claim 1, wherein the plurality of stored audio patterns are frequency domain representations of the spoken user input.

12. The method of claim 6, wherein:

a processor is operable to convert a discrete time signal representing audio input and a discrete time signal representing an audio specimen to the frequency domain to create a frequency domain representation of the audio input and a frequency domain representation of the audio specimen; and

determining the highest correspondence between the stored audio pattern and the digital audio input comprises calculating a correlation between the frequency domain representation of the digital audio input and each of the plurality of frequency domain representations of the stored audio patterns.

13. The method of claim 2, wherein the output signal is generated in a central hub and is distributed to one or more peripheral devices operably connected to the central hub.

14. The method of claim 6, wherein the table compares the stored audio patterns associated with the user with digital audio input signals if the digital audio input signals contain audio input under a predetermined time slice duration threshold;

and

the table does not compare the digital audio input signals with stored audio patterns if the audio input in a digital audio input signal is above a predetermined time slice duration threshold.

20. The method of claim 3, wherein the stored audio patterns and digital tokens are stored in a table and one of the plurality of stored audio patterns in the table are selected by:

comparing a digital audio input signal with each stored audio pattern associated with the identified user; and

selecting one of the plurality of stored audio patterns above a specified threshold of correspondence with the highest correspondence to the digital audio input.

15. A portable voice recognition apparatus for generating an output signal based on spoken user input comprising:

an acoustic-electric transducer configured to receive monosyllabic audio input and output an electrical audio input signal;

an analog to digital converter operably connected to the acoustic-electric transducer configured to take the electrical audio input signal and output a discrete electrical audio input signal;

memory including;

a plurality of stored audio patterns generated by the user, each audio pattern including a single monosyllabic representation of a command, number, letter or word having two or more syllables in a custom language wherein at least one audio pattern represents both an opening command and a closing command; and

a plurality of digital tokens each associated with at least one stored audio pattern;

at least one processor operably connected to the at least one memory wherein the processor is configured to;

receive the electrical audio input signal;

determine the stored audio pattern associated with the electrical audio input signal;

determine the digital token associated with the selected one of the plurality of stored audio patterns; and

generate an output signal configured to communicate the command, number, letter or word based on the determined digital token;

at least one output control circuit; and

at least one output port operably connected to the at least one output control circuit.

16. The portable voice recognition apparatus of claim 15, wherein each of the plurality of stored audio patterns is associated with one distinct monosyllabic audio input.

17. The portable voice recognition apparatus of claim 15, further comprising:

a threshold determining component operably connected to the memory to determine if the discrete audio input signal contains audio input under a predetermined time slice duration.

18. A portable voice recognition apparatus for generating an output signal based on spoken user input comprising:

memory including custom language entries made up of custom language elements wherein each element includes a single monosyllabic representation of a command, number, letter or word having two or more syllables wherein at least one custom language element represents both an opening command and a closing command;

at least one of an audio port and visual display means to communicate to a user the entire set of custom language elements stored in memory;

an acoustic-electric transducer to convert each uttered monosyllabic audio specimen received from the user into a continuous electrical signal;

an analog to digital converter to convert each continuous electrical signal into a discrete electrical signal;

a processor for conditioning each discrete electrical signal; and

a look up table, stored in the memory, for storing each conditioned discrete electrical signal as a stored audio pattern.

19. The portable voice recognition apparatus of claim 18, wherein each of the stored audio patterns is associated with one distinct monosyllabic audio input.