Speech recognition system for an automotive vehicle

• US 4,538,295 A
• Filed: 08/16/1982
• Issued: 08/27/1985
• Est. Priority Date: 08/16/1982
• Status: Expired due to Fees

• Current Assignee: Nissan Motor Co., Ltd.
• Original Assignee: Nissan Motor Co., Ltd.
• Inventors: Kishi, Norimasa, Noso, Kazunori
• Primary Examiner(s): Kemeny, E. S. Matt
• Application Number:
  US06/408,699
• Time in Patent Office: 1,107 Days
• Field of Search: 381/46, 381/47, 381/71, 381/86, 455/297, 455/312
• US Class Current: 704/248
• CPC Class Codes: B60R 16/0373 Voice control in general G10L
• CPC Class Codes: G10L 15/20 Speech recognition techniques specially adapted for robustness in adverse environments, e.g. in noise, of stress induced speech G10L21/02 takes precedence
• CPC Class Codes: G10L 25/87 Detection of discrete points within a voice signal

Abstract

A speech recognition system for an automotive vehicle can effectively eliminate noise generated by vehicle engines, and comprises a group of bandpass filters the pass bands of which are approximately from 500 to 600, from 900 to 1200, and from 900 to 2200 Hz, for effectively passing only the speech sound formant frequencies while eliminating the major frequency components of engine noise less than 500 Hz, in addition to the conventional speech recognizer. Further, it is desirable to provide the speech recognition system with a high-frequency compensator for enhancing the voltage level of the high frequency components of the electrical signal representing spoken instructions at a rate of 20 dB or more per decade increase in frequency and/or a plurality of notch filters fo sharply reducing the major engine noise frequency components at 200, 400, and 800 Hz.

Claims

1. A speech recognition system for an automotive vehicle for activating vehicle actuators in response to a spoken instruction signal received through a microphone, comprising:

(a) a voice input section including notch filter means for attenuating frequency components of the signals received through the microphone at approximately 400 Hz;

(b) voice detection means connected to said notch filter means for detecting beginning and end of spoken instruction signals and outputting instruction start and end signals for identifying, respectively, start and end of an instruction;

(c) a plurality of bandpass filter means connected to said notch filter means for further filtering signals filtered through said notch filter means to human speech frequency components and to noise frequency components due to noise resulting from engine operation; and

(d) voice analysis means connected to said bandpass filter means and responsive to the instruction start and end signals for analyzing the signals filtered through said bandpass filter means, comparing analyzed results with predetermined reference values corresponding to at least one spoken instruction, and for activating at least one actuator when the analyzed results match predetermined reference values associated with the actuator.

2. A speech recognition system for an automotive vehicle as set forth in claim 1, wherein the lowest central passband frequency of said bandpass filter means is above 500 Hz.

3. A speech recognition system for an automotive vehicle as set forth in claim 1, wherein passband frequencies of said bandpass filter means range from 500 to 600 Hz and from 900 to 2200 Hz.

4. A speech recognition system for an automotive vehicle as set forth in claim 1, wherein the passband frequencies of said bandpass filter means range from 500 to 600 Hz and 900 to 3800 Hz.

6. The speech recognition system as set forth in claim 1, wherein said notch filter means further attentuates frequency components of the signals form the microphone at approximately 200 Hz and at approximately 800 Hz.

11. The speech recognition system as set forth in claim 2, wherein said notch filter means further attenuates, frequency components of the signals from the microphone at approximately 200 Hz and at approximately 800 Hz.

12. A speech recognition system for an automotive vehicle as set forth in claim 4, wherein passband frequencies of said bandpass filter means range from 500 to 600 Hz and 900 to 3800 Hz.

5. In a speech recognition system for automotive vehicle for activating vehicle actuators in response to spoken instructions received through a microphone, having

(a) a sound input section for filtering signals from a microphone,

(b) a voice detection section connected to said sound input section for detecting the beginning and end of spoken instructions and for outputting instruction start and end signals for identifying, respectively, start and end of an instruction, and

(c) a voice analysis section connected to said input section and responsive to said instruction start and end signals, the improvement comprising:

compensating means in said sound input section for compensating for variable frequency characteristic noise signals present in said signals from a microphone, said compensating means including filter means comprised of first means having a frequency response characterized by a high frequency characteristic substantially inverse to a high frequency characteristic of said noise signals for cancelling said noise signals and for providing a substantially flat response at said high frequencies to said noise signals.

7. The speech recognition system as set forth in claim 5, wherein said high-frequency compensating means amplifies frequency components of the signal from the microphone to a degree which increases with frequency at a rate of 20 dB or more per decade.

8. The speech recognition system as set forth in claim 7, wherein said high-frequency compensator amplifies frequency components of the signal from the microphone lower than about 200 Hz to a degree which decreases with frequency, so that the frequency components lower than 200 Hz are attenuated in comparison to other frequencies, at a rate of -20 dB or less per decade.

9. A speech recognition system as set forth in claim 5 wherein said filter means includes second means characterized by a frequency response having a predetermined low frequency characteristic substantially inverse to a low frequency characteristic of said noise signals for cancelling said noise signals and for providing a substantially flat response at said high frequencies and said low frequencies to said noise signals.

10. A speech recognition system as set forth in claim 5 wherein said filter means is characterized by a negative slope low frequency response and by a positive slope high frequency response for providing a substantially flat response to said noise signal at both high and low frequencies.