System and method for improved speech acquisition for hands-free voice telecommunication in a noisy environment

• US 5,353,376 A
• Filed: 03/20/1992
• Issued: 10/04/1994
• Est. Priority Date: 03/20/1992
• Status: Expired due to Term

• Current Assignee: Texas Instruments, Inc.
• Original Assignee: Texas Instruments, Inc.
• Inventors: Viswanathan, Vishu R., Oh, Sang G.
• Primary Examiner(s): Fleming, Michael R.
• Assistant Examiner(s): KIM, RICHARD
• Application Number:
  US07/854,161
• Time in Patent Office: 928 Days
• Field of Search: 381/41, 381/46, 395/2.42
• US Class Current: 704/233
• 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 2021/02166 Microphone arrays; Beamforming
• CPC Class Codes: G10L 21/0208 Noise filtering

Abstract

Systems and methods for improved speech acquisition are disclosed including a plurality of linearly arrayed sensors to detect spoken input and to output signals in response thereto, a beamformer connected to the sensors to cancel a preselected noise portion of the signals to thereby produce a processed signal, and a speech recognition system to recognize the processed signal and to respond thereto. The beamformer may also include an adaptive filter with enable/disable circuitry for selectively training the adaptive filter a predetermined period of time. A highpass filter may also be used to filter a preselected noise portion of the sensed signals before the signals are forwarded to the beamformer. The speech recognition system may include a speaker independent base which is able to be adapted by a predetermined amount of training by a speaker, and which system includes a voice dialer or a speech coder for telecommunication.

Claims

1. A system for an improved acquisition of speech comprising:

a plurality of linearly arrayed sensors to detect spoken input and to output signals in response thereto;

a highpass filter to filter a preselected first noise portion from said signals received from said linearly arrayed sensors;

a beamformer, connected to said highpass filter, to cancel a preselected second noise portion from said filtered signals thereby producing a processed signal; and

a speech application system, connected to receive said processed signal from said beamformer, for receiving and responding to said processed signal, wherein said beamformer is a generalized sidelobe canceler, comprising a delay and sum beamformer, an adaptive noise canceler and an adder which produces said processed signal,

and wherein said adaptive noise canceler comprises a blocking matrix, an adaptive filter and an adder, and wherein said adaptive filter comprises enable/disable circuitry for selectively adapting said adaptive filter for a predetermined period of time.

2. A system for an improved acquisition of speech comprising:

a plurality of linearly arrayed sensors to detect spoken input and to output signals in response thereto;

a highpass filter to filter a preselected first noise portion from said signals received from said linearly arrayed sensors;

a beamformer, connected to said highpass filter, to cancel a preselected second noise portion from said filtered signals thereby producing a processed signal; and

a speech application system, connected to receive said processed signal from said beamformer, for receiving and responding to said processed signal, wherein said speech application system is a speech coder with a communications link.

3. A system for an improved acquisition of speech comprising:

a plurality of linearly arrayed sensors to detect spoken input and to output signals in response thereto;

a highpass filter to filter a preselected first noise portion from said signals received from said linearly arrayed sensors;

a beamformer, connected to said highpass filter, to cancel a preselected second noise portion from said filtered signals thereby producing a processed signal; and

a speech application system, connected to receive said processed signal from said beamformer, for receiving and responding to said processed signal, wherein said speech application system is a speech recognition system.

4. The system of claim 3, wherein said speech recognition system includes a speaker recognition algorithm which is adaptable by a predetermined amount of training by a speaker.

5. The system of claim 4, wherein said speaker recognition algorithm is speaker independent.

6. The system of claim 3, wherein said speech recognition system uses a Hidden Markov model speech recognition algorithm.

7. A system for improved acquisition of speech, comprising of:

a plurality of linearly arrayed sensors to detect spoken input and to output signals in response thereto;

a beamformer, connected to said linearly arrayed sensors, to cancel a preselected noise portion of said signals thereby producing a processed signal, said beamformer including an adaptive noise canceler and an adder which produces said processed signal; and

a speech application system, connected to receive said processed signal from said beamformer, for receiving and responding to said processed signal, wherein said adaptive noise canceler comprises a blocking matrix, an adaptive filter and an adder, and wherein said adaptive filter further comprises enable/disable circuitry for selectively adapting said adaptive filter for a predetermined period of time.

8. A system for improved acquisition of speech, comprising of:

a plurality of linearly arrayed sensors to detect spoken input and to output signals in response thereto;

a beamformer, connected to said linearly arrayed sensors, to cancel a preselected noise portion of said signals thereby producing a processed signal, said beamformer including an adaptive noise canceler and an adder which produces said processed signal; and

a speech application system, connected to receive said processed signal from said beamformer, for receiving and responding to said processed signal, wherein said speech application system includes a speech coder and a communications link.

9. A system for improved acquisition of speech, comprising of:

a plurality of linearly arrayed sensors to detect spoken input and to output signals in response thereto;

a beamformer, connected to said linearly arrayed sensors, to cancel a preselected noise portion of said signals thereby producing a processed signal, said beamformer including an adaptive noise canceler and an adder which produces said processed signal; and

a speech application system, connected to receive said processed signal from said beamformer, for receiving and responding to said processed signal, wherein said speech application system is a speech recognition system.

10. The system of claim 9, wherein said speech recognition system includes a speaker recognition algorithm which is adaptable by a predetermined amount of training by a speaker.

11. The system of claim 10, wherein said speaker recognition algorithm is speaker independent.

12. The system of claim 9, wherein said speech recognition system uses a Hidden Markov Model speech recognition algorithm.

13. A system for improved acquisition of speech, comprising:

a plurality of linearly arrayed sensors to detect spoken input and to output signals in response thereto;

a beamformer, connected to said linearly arrayed sensors, to cancel a preselected noise portion of said signals thereby producing a processed signal; and

a speech recognition system, connected to receive said processed signal from said beamformer, for receiving and responding to said processed signal, said speech recognition system including a speaker recognition algorithm which is adaptable by a predetermined amount of training by a speaker.

14. They system of claim 13, wherein said system further comprises a highpass filter connected to said beamformer to filter a preselected noise portion of said signals received from said linearly arrayed sensors before transmitting such filtered signals to said beamformer.

15. The system of claim 13, wherein said beamformer is a generalized sidelobe canceler, comprising a delay and sum beamformer, an adaptive noise canceler and an adder which produces said processed signal.

16. The system of claim 15, wherein said adaptive noise canceler comprises a blocking matrix, an adaptive filter and an adder.

17. The system of claim 16, wherein said adaptive filter further comprises enble/disable circuitry for selectively adapting said adaptive filter for a predetermined period of time.

18. The system of claim 13, wherein said speech recognition system uses a Hidden Markov Model speech recognition algorithm.

19. A method for improved acquisition of speech during hands-free cellular telephony, comprising the steps of:

detecting spoken signals, including a speech portion and a noise portion, with linearly arrayed sensors;

filtering a preselected portion of said noise portion of said signals by a highpass filter;

processing said filtered signals with a beamformer including an adaptive filter and enable/disable circuitry for selectively adapting said adaptive filter for a predetermined period of time;

recognize said speech portion; and

responding to said processed signal.

20. The method of claim 19, wherein said step of responding to said processed signal includes the step of dialing a number associated with said speech portion.

21. The method of claim 19, wherein said step of recognizing said speech portion further includes the step of employing a Hidden Markov Model speech recognition algorithm to aid in recognizing such speech portion.

22. The method of claim 19, wherein the method further comprises the step of adapting an adaptive filter for a predetermined period of time and the step of stopping said adaptation of said adaptive filter after said step of detecting step.