Adaptive biofeedback speech tutor toy

• US 5,562,453 A
• Filed: 02/02/1993
• Issued: 10/08/1996
• Est. Priority Date: 02/02/1993
• Status: Expired due to Fees

• Current Assignee: Sheree H.-R. Wen
• Original Assignee: Sheree H.-R. Wen
• Inventors: Wen, Sheree H.-R.
• Primary Examiner(s): Apley, Richard J.
• Assistant Examiner(s): Richman, Glenn E.
• Application Number:
  US08/016,981
• Time in Patent Office: 1,344 Days
• Field of Search: 434/185, 434/156, 434/157, 434/167, 434/169, 434/176, 434/178, 364/419.03, 395/2, 381/36-43, 381/45
• US Class Current: 434/185
• CPC Class Codes: G10L 15/22 Procedures used during a speech recognition process, e.g. man-machine dialogue
• CPC Class Codes: G10L 2015/221 Announcement of recognition results

Abstract

A prompt-providing adaptive speech tutor toy, is characterized by a programmed sequence of prompts in standard speech evoking a user'"'"'s voice response of a word or phrase. Initially, a vocalized password provides a wake-up call during sleep mode to activate a start sequencer. This turns on the system. The user selects a category. A sequence of prompts and responses follows. Each response is accepted by a user input transducer, electronically treated to eliminate environmental noise and to put the response in a digitized form for computer evaluation. Computer evaluation follows. Frequency domain information from the user pattern is compared against frequency domain information from a standard pattern. Time domain information from the user pattern is compared against time domain information from a standard pattern. The comparisons are concatenated into a comparison score and this is checked against a threshold. Overthreshold score triggers applause biofeedback and a new prompt; underthreshold score triggers a re-prompt. Various mode changes may be made for mimic response, Q&A response, mathematics, music or others. Voice turn-on and initialization for the voice characteristics of the toy user are automatic. The pattern comparator is altered for optimum match to the frequency clustering pattern of the user, and where appropriate the standard pattern is altered for better match to the toy user'"'"'s voice.

Claims

1. A prompt-providing adaptive biofeedback speech tutor system for voice response to a programmed sequence of prompts and related responses by the user, characterized by:

a) a user input transducer (2);

b) an audible transducer (6);

c) memory means (8,

     9) including partition memory respectively for frequency domain and time domain information;

d) a microchannel analyzer (7) connected to said user input transducer (2) for input so as to provide outputs to said memory means (8,

     9);

e) pattern comparator means (10) having memory means for frequency domain information of a user pattern against frequency domain information of a standard pattern and also of a time domain user pattern against a time domain standard pattern, to achieve a comparison score;

f) means responsive to the comparison score in said pattern comparator means (10) to calculate overthreshold and underthreshold values and to provide respectively related overthreshold and underthreshold signals;

f) biofeedback configuring means responsive to said overthreshold signal; and

g) sequencing means responsive to said overthreshold signal and to said underthreshold signal to control the sequence of prompts.

2. A prompt-providing adaptive biofeedback speech tutor system according to claim 1, wherein said input transducer means (2, 3) includes both microphone (2) and voice activated start sequencer (3).

3. A prompt-providing adaptive biofeedback speech tutor system according to claim 2, wherein said input transducer means (2,3,4) includes microphone (2), voice activated start sequencer (3) and environmental noise filter (4).

4. A prompt-providing adaptive biofeedback speech tutor system according to claim 1, wherein said input transducer means (2,4) includes microphone (2) and environmental noise filter (4).

5. A prompt-providing adaptive biofeedback speech tutor system according to claim 1, wherein said pattern comparator means (10) includes means to provide a level score which is a composite of a multiplicity of individual comparison values.

6. A prompt-providing adaptive biofeedback speech tutor system according to claim 1, wherein said pattern comparator means (10) includes means to compare said level score with a threshold score and provide a composite overthreshold/underthreshold output.

7. A prompt-providing adaptive biofeedback speech tutor system according to claim 5, wherein said pattern comparator means (10) includes means responsive to said composite score overthreshold/underthreshold output to alter the next prompt.

8. A prompt-providing adaptive biofeedback speech tutor system according to claim 7, wherein said means to alter the next prompt selects a variation of the previous prompt with emphasis which is a function of a pattern of at least one of the individual comparison values making up the composite score underthreshold/overthreshold output.

9. A prompt-providing adaptive biofeedback speech tutor system according to claim 1, wherein the user responds simultaneously with the prompt and said evaluation means includes means to eliminate the effect of the prompt as picked up by said microphone (2).

10. A prompt-providing adaptive biofeedback speech tutor system according to claim 1, wherein the system has a sleep mode during which a static read only memory is in constant readiness for a password wake-up call, comprising, in addition, means to provide a simple overthreshold determination with respect to a voice wake-up call and turn on the toy in response to such overthreshold.

11. A prompt-providing adaptive biofeedback speech tutor system according to claim 10, wherein the system in response to said wake-up call controls an initialization sequence to optimize frequency clustering as a result of analysis of the password of the wake-up call.