Speech dialogue system for realizing improved communication between user and system

• US 5,548,681 A
• Filed: 08/13/1992
• Issued: 08/20/1996
• Est. Priority Date: 08/13/1991
• Status: Expired due to Term

• Current Assignee: Kabushiki Kaisha Toshiba (Toshiba Corporation)
• Original Assignee: Kabushiki Kaisha Toshiba (Toshiba Corporation)
• Inventors: Gleaves, David, Takebayashi, Yoichi, Nagata, Yoshifumi
• Primary Examiner(s): Knepper, David D.
• Application Number:
  US07/929,106
• Time in Patent Office: 1,468 Days
• Field of Search: 395/2.35-2.37, 395/2.42, 395/2.4, 395/2, 395/2.65, 381/41-43, 381/46, 381/47
• US Class Current: 704/233
• CPC Class Codes: G10L 15/22 Procedures used during a speech recognition process, e.g. man-machine dialogue
• CPC Class Codes: G10L 2021/02087 the noise being separate speech, e.g. cocktail party

Abstract

In the system, a speech input uttered by a human is received by a microphone which outputs microphone output signals. The speech input received by the microphone is then recognized by a speech recognition unit, and a synthetic speech response appropriate for the speech input recognized by the speech recognition unit is generated and outputted from a loudspeaker to the human. In recognizing the speech input, the speech recognition unit receives input signals in which the synthetic speech response, outputted from the loudspeaker and then received by the microphone, is cancelled from the microphone output signals.

Claims

1. A speech dialogue system, comprising:

microphone means for receiving a speech input uttered by a human speaker and outputting microphone output signals;

speech recognition means for receiving input signals and recognizing the speech input received by the microphone means;

synthetic speech response generation means for generating a synthetic speech response appropriate for the speech input recognized by the speech recognition means;

loudspeaker means for outputting the synthetic speech response to the human speaker; and

synthetic speech response cancellation means for cancelling the synthetic speech response, which is outputted from the loudspeaker means and then received by the microphone means, from the microphone output signals, to obtain input signals to be supplied to the speech recognition means from which the speech recognition means recognizes the speech input, the synthetic speech response cancellation means further comprising;

look-up table means for memorizing speech characteristic information on the synthetic speech response to be outputted from the loudspeaker means;

adaptive filter means for adapting the synthetic speech response generated by the synthetic speech response generation means by multiplying the synthetic speech response with filter coefficients calculated according to the speech characteristic information memorized in the look-up table means, to obtain an adapted synthetic speech response; and

subtractor means for subtracting the adapted synthetic speech response obtained by the adaptive filter means from the microphone output signals, to obtain the input signals supplied to the speech recognition means.

2. The speech dialogue system of claim 1, wherein the speech characteristic information memorized in the look-up table means indicates an auto-correlation characteristic of the synthetic speech response.

3. The speech dialogue system of claim 1, wherein the speech characteristic information memorized in the look-up table means indicates a power characteristic of the synthetic speech response.

4. The speech dialogue system of claim 1, further comprising control means for controlling an updating of the filter coefficients used in the adaptive filter means according to a presence of the speech input recognized by the speech recognition means.

5. The speech dialogue system of claim 4, wherein the control means sets values of the filter coefficients in the adaptive filter to values before the utterance of the speech input when the presence of the speech input is recognized by the speech recognition means.

6. The speech dialogue system of claim 5, wherein the control means controls the updating of the filter coefficients used in the adaptive filter means such that the filter coefficients are not changed unless a power level of the speech input is greater than a predetermined threshold.

7. The speech dialogue system of claim 1, further comprising speech segment detection means for detecting a speech segment in the input signals, wherein the speech recognition means recognize the speech input from the input signals only at the speech segment detected by the speech segment detection means.

8. The speech dialogue system of claim 7, wherein the speech segment detection means uses detection thresholds for determining the speech segment which vary according to a power level of the synthetic speech response to be cancelled at the synthetic speech response cancellation means.

9. The speech dialogue system of claim 8, wherein the speech segment detection means obtains the power level of the synthetic speech response by calculating a convolution of the synthetic speech response and an estimate for the filter coefficients.

10. The speech dialogue system of claim 9, wherein the estimate for the filter coefficients is obtained by adding a wide frequency band noise to the synthetic speech response.

11. The speech dialogue system of claim 9, wherein the estimate for the filter coefficients is obtained by a spectral pre-whitening of a frequency spectrum of the synthetic speech response.

13. The speech dialogue system of claim 7, wherein the smoothing filter means includes a first smoothing filter having a first time constant and a second smoothing filter having a second time constant which is larger than the first time constant, and the switching means stops the operation of the adaptive filter means when the output power level of the first smoothing filter is below a predetermined first threshold, and starts the operation of the adaptive filter means when the output power level of the second smoothing filter is above a predetermined second threshold.

14. The speech dialogue system of claim 13, wherein the filter coefficients used in the adaptive filter means are updated according to the output power level of the second smoothing filter.

12. A speech dialogue system, comprising:

microphone means for receiving a speech input uttered by a human speaker and outputting microphone output signals;

speech recognition means for receiving input signals and recognizing the speech input received by the microphone means;

synthetic speech response generation means for generating a synthetic speech response appropriate for the speech input recognized by the speech recognition means;

loudspeaker means for outputting the synthetic speech response to the human speaker; and

synthetic speech response cancellation means for cancelling the synthetic speech response, which is outputted from the loudspeaker means and then received by the microphone means, from the microphone output signals, to obtain input signals to be supplied to the speech recognition means from which the speech recognition means recognizes the speech input, the synthetic speech response cancellation means further comprising;

adaptive filter means for adapting the synthetic speech response generated by the synthetic speech response generation means, to obtain an adaptive filter output;

convolution means for calculating a convolution of the adaptive filter output obtained by the adaptive filter means and the synthetic speech response generated by the synthetic speech response generation means;

subtractor means for subtracting the convolution calculated by the convolution means from the microphone output signals, to obtain the input signals supplied to the speech recognition means;

at least one smoothing filter means for smoothing the synthetic speech response generated by the synthetic speech response generation means; and

switching means for controlling an operation of the adaptive filter means according to an output power level of the smoothing filter means.

15. The speech dialogue system of claim 12, wherein the synthetic speech response cancellation means further comprises two channel A/D converter means for inputting the synthetic speech response generated by the synthetic speech response generation means and the microphone output signals outputted by the microphone means into the synthetic speech response cancellation means synchronously.

16. The speech dialogue system of claim 12, further comprising speech segment detection means for detecting a speech segment in the input signals, wherein the speech recognition means recognize the speech input from the input signals only at the speech segment detected by the speech segment detection means.