Interactive speech recognition with varying responses for time of day and environmental conditions

• US 5,802,488 A
• Filed: 02/29/1996
• Issued: 09/01/1998
• Est. Priority Date: 03/01/1995
• Status: Expired due to Term

• Current Assignee: Seiko Epson Corporation (Seiko Group)
• Original Assignee: Seiko Epson Corporation (Seiko Group)
• Inventors: Edatsune, Isao
• Primary Examiner(s): Knepper, David D.
• Application Number:
  US08/609,336
• Time in Patent Office: 915 Days
• Field of Search: 395/2.4, 395/2.41, 395/2.45-2.49, 395/2.5, 395/2.6, 395/2.67, 395/2.79, 395/2.81, 395/2.83, 704/231, 704/232, 704/236-239, 704/240, 704/241, 704/251, 704/258, 704/270, 704/272, 704/275, 704/274, 704/257
• US Class Current: 704/231
• CPC Class Codes: A63H 2200/00 Computerized interactive toys, e.g. dolls
• CPC Class Codes: G10L 13/00 Speech synthesis; Text to speech systems
• CPC Class Codes: G10L 15/26 Speech to text systems G10L15/08 takes precedence
• CPC Class Codes: G10L 2015/226 using non-speech characteristics

Abstract

The invention improves recognition rates by providing an interactive speech recognition device that performs recognition by taking situational and environmental changes into consideration, thus enabling interactions that correspond to situational and environmental changes. The invention comprises a speech analysis unit that creates a speech data pattern corresponding to the input speech; a timing circuit for generating time data, for example, as variable data; a coefficient setting unit receiving the time data from the timing circuit and generating weighting coefficients that change over time, in correspondence to the content of each recognition target speech; a speech recognition unit that receives the speech data pattern of the input speech from the speech analysis unit, and that at the same time obtains a weighting coefficient in effect for a pre-registered recognition target speech at the time from the coefficient setting unit, that computes final recognition data by multiplying the recognition data corresponding to each recognition target speech by its corresponding weighting coefficient, and that recognizes the input speech based on the computed final recognition result; a speech synthesis unit for outputting speech synthesis data based on the recognition data that takes the weighting coefficient into consideration; and a drive control unit for transmitting the output from the speech synthesis unit to the outside.

Claims

1. An interactive speech recognition device, comprising:

speech analysis means for analyzing an input speech and creating a speech data pattern that matches characteristics of the input speech;

detection means for detecting variable non-speech data that changes speech flowing from the speech recognition device;

coefficient setting means, responsive to the variable non-speech data, for generating a plurality of weighting coefficients each pre-assigned to a pre-registered recognition target speech, based on the variable non-speech data;

speech recognition means for computing a final recognition result in response to the speech data pattern, said speech recognition means including;

means for storing a plurality of pre-registered recognition target speeches and for outputting, in response to the speech data pattern, a plurality of recognition data values each for a corresponding pre-registered recognition target speech,

means for computing final recognition data by multiplying each recognition data value by a corresponding one of said pre-assigned weighting coefficients for a corresponding pre-registered recognition target speech, and

means for recognizing the input speech by comparing the final recognition data for all of the pre-registered recognition target speeches and for outputting a final recognition result; and

speech synthesis means for converting the final recognition result to corresponding speech synthesis data for producing an appropriate response to the input speech.

2. The interactive speech recognition device of claim 1, wherein said detection means includes timing means for providing time data, and each of the weighting coefficients generated by said coefficient setting means corresponds to the time data of a day for a corresponding pre-registered recognition target speech.

3. The interactive speech recognition device of claim 2, further comprising coefficient storage means, responsive to the time data from said timing means, for storing past time data relating to past statistic data and for creating weighting coefficients based on the past time data relating to the past statistic data, wherein said coefficient setting means, responsive to said timing means and said coefficient storage means, generates a preset largest value of a weighting coefficient for a pre-selected, pre-registered recognition target speech if the input speech occurs at a peak time at which it was correctly recognized most frequently in the past, and generates a smaller value of the weighting coefficient as time deviates from this peak time.

4. An interactive speech recognition device, comprising:

speech analysis means for analyzing an input speech and creating a speech data pattern that matches characteristics of the input speech;

speech recognition means for generating recognition data that corresponds to the input speech based on the speech data pattern created by said speech analysis means;

timing means for generating time data;

response content level storage means for storing information relating to passage of time relative to a response content level;

response content level generation means for storing time ranges for a plurality of response content levels, said response content level generation means being responsive to the time data from said timing means, the recognition data from said speech recognition means, and the information from said response content level storage means, for generating a response content level value according to passage of time;

response content creation means, responsive to the recognition data from said speech recognition means and the response content level value from said response content level generation means, for determining response content data appropriate for the response content level value generated by said response content level generation means; and

speech synthesis means for converting the response content data from said response content creation means to corresponding speech synthesis data for producing an appropriate response to the input speech.

5. An interactive speech recognition device, comprising;

speech analysis means for analyzing an input speech and creating a speech data pattern that matches characteristics of the input speech;

speech recognition means for generating recognition data that corresponds to the input speech, based on the speech data pattern from said speech analysis means;

detection means for detecting variable non-speech data that changes speech flowing from the speech recognition device;

response content creation means, responsive to the variable non-speech data from said detection means and the recognition data from said speech recognition means, for outputting response content data, based on the recognition data by taking the variable non-speech data into consideration,

speech synthesis means for converting the response content data to corresponding speech synthesis data for producing an appropriate response to the input speech.

6. The interactive speech recognition device of claim 5, wherein said detection means includes a temperature sensor that measures an environmental temperature and outputs temperature data, and said response content creation means outputs the response content data by taking the temperature data into consideration.

7. The interactive speech recognition device of claim 5, wherein said detection means includes an air pressure sensor that measures an environmental air pressure and outputs air pressure data, and said response content creation means outputs the response content data by taking the air pressure data into consideration.

8. The interactive speech recognition device of claim 5, wherein said detection means includes calendar detection means for detecting calendar data and outputting the calendar data, and said response content creation means outputs the response content data by taking the calendar data into consideration.

9. The interactive speech recognition device of claim 5, wherein said detection means includes timing means for providing time data, and response content data generated by said response content creation means corresponds to the time data of a day for a corresponding pre-registered recognition target speech.

10. An interactive speech recognition device, comprising:

speech analysis means for analyzing an input speech and creating a speech data pattern that matches characteristics of the input speech;

detection means for detecting time data that changes speech flowing from the speech recognition device;

coefficient setting means, responsive to the detected time data, for generating a plurality of weighting coefficients each pre-assigned to a pre-registered recognition target speech, based on the time data;

speech recognition means for computing a final recognition result in response to the speech data pattern, said speech recognition means including;

means for storing a plurality of pre-registered recognition target speeches and for outputting, in response to the speech data pattern, a plurality of recognition data values each for a corresponding pre-registered recognition target speech,

means for computing final recognition data by multiplying each recognition data value by a corresponding one of said pre-assigned weighting coefficient for a corresponding pre-registered recognition target speech, and

means for recognizing the input speech by comparing the final recognition data for all of the pre-registered recognition target speeches and for outputting a final recognition result; and

speech synthesis means for converting the final recognition result to corresponding speech synthesis data for producing an appropriate response to the input speech.

11. An interactive speech recognition device, comprising:

speech analysis means for analyzing an input speech and creating a speech data pattern that matches characteristics of the input speech;

speech recognition means for generating recognition data that corresponds to the input speech, based on the speech data pattern from said speech analysis means;

detection means for detecting variable non-speech data that changes speech flowing from the speech recognition device;

response content creation means, responsive to the variable non-speech data from said detection means and the recognition data from said speech recognition means, for outputting response content data, based on the recognition data by taking the variable data into consideration;

an operating mechanism and a drive control unit responsive to the response content data for controlling the operating mechanism.

12. The interactive speech recognition device of claim 11 further comprising speech synthesis means for converting said response content data to speech synthesis data simultaneous with said drive control unit controlling said operating mechanism in response to said response content data.