Robust adaptive beamforming with enhanced noise suppression

• US 8,818,002 B2
• Filed: 07/21/2011
• Issued: 08/26/2014
• Est. Priority Date: 03/22/2007
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Tashev, Ivan, Acero, Alejandro, Yoon, Byung-Jun
• Primary Examiner(s): Saunders, Jr., Joseph
• Assistant Examiner(s): MOONEY, JAMES K
• Application Number:
  US13/187,618
• Publication Number:
  US 20110274291A1
• Time in Patent Office: 1,132 Days
• Field of Search: None
• US Class Current: 381/94.2
• CPC Class Codes: G01S 3/86 with means for eliminating undesired waves, e.g. disturbing noises
• CPC Class Codes: H04B 7/0854 using error minimizing algorithms, e.g. minimum mean squared error [MMSE], "cross-correlation" or matrix inversion
• CPC Class Codes: H04R 2430/20 Processing of the output signals of the acoustic transducers of an array for obtaining a desired directivity characteristic H04R2203/12 takes precedence
• CPC Class Codes: H04R 3/005 for combining the signals of two or more microphones specially adapted for hearing aids H04R25/407

Abstract

A novel adaptive beamforming technique with enhanced noise suppression capability. The technique incorporates the sound-source presence probability into an adaptive blocking matrix. In one embodiment the sound-source presence probability is estimated based on the instantaneous direction of arrival of the input signals and voice activity detection. The technique guarantees robustness to steering vector errors without imposing ad hoc constraints on the adaptive filter coefficients. It can provide good suppression performance for both directional interference signals as well as isotropic ambient noise.

Claims

1. A system for improving the signal to noise ratio of a desired signal received from a microphone array relative to at least one undesired signal received from the microphone array, comprising:

a general purpose computing device;

a computer program comprising program modules executable by the general purpose computing device, wherein the computing device is directed by the program modules of the computer program to,

input signals of sensors of a sensor array in the frequency domain defined by frequency bins and a time frame index, wherein the input signals comprise a desired signal and at least one undesired signal;

compute a beamformer output as function of the input signals divided into k frequency bins and the time frame index;

estimate the desired signal by taking the product of the beamformer output and the probability that the beamformer output contains the desired signal at a given target angle region;

use the estimated desired signal as a reference signal and subtracting components that are correlated to the estimated desired signal from the input signals to obtain interference signals;

subtract the interference signals from the beamformer output to produce an output signal of the desired signal with an increased signal to noise ratio.

2. The system of claim 1 wherein at least one undesired signal has a magnitude greater than the desired signal.

3. The system of claim 1 wherein the input signals of the sensor array in the frequency domain are converted from the time domain into the frequency domain prior to inputting them using a Modulated Complex Lapped Transform (MCLT).

4. The system of claim 1 wherein the sensors are microphones and wherein the sensor array is a microphone array.

5. The system of claim 1 wherein the probability that the beamformer output contains the desired signal at a given target angle region is computed by multiplying the probability that a signal component in a k-th frequency bin originates inside a specified tracking region by the probability that the desired signal exists in the k-th frequency bin.

6. The system of claim 5 wherein the specified tracking region is defined by the expected direction of arrival of the desired signal and a maximum allowed direction of arrival error.

7. The system of claim 5 wherein the probability that a signal component in the k-th frequency bin originates inside a specified tracking region employs an instantaneous direction of arrival for each frequency bin based on the phase differences of the input signals.

8. The system of claim 5 wherein the probability that the desired signal exists in the k-th frequency bin is determined using a voice activity detector.

9. The system of claim 1 wherein the system comprises an electronic consumer device.

10. A computer-implemented process for cancelling undesired sounds from a signal received at a microphone array, comprising:

inputting signals of sensors of a sensor array in the frequency domain defined by frequency bins and a time frame index, wherein the input signals comprise a desired signal and at least one undesired signal;

computing a beamformer output as function of the input signals divided into k frequency bins and the time frame index;

estimating the desired signal by taking the product of the beamformer output and the probability that the beamformer output contains the desired signal at a given target angle region;

using the estimated desired signal as a reference signal and subtracting components that are correlated to the estimated desired signal from the input signals to obtain interference signals;

subtracting the interference signals from the beamformer output to produce an output signal of the desired signal with an increased signal to noise ratio while attenuating any undesired signal.

11. The computer-implemented process of claim 10 wherein the desired signal does not have a magnitude as large as an undesired signal.

12. The computer-implemented process of claim 10 wherein the probability that the beamformer output contains the desired signal at a given target angle region is computed by multiplying the probability that a signal component in a k-th frequency bin originates inside a specified tracking region by the probability that the desired signal exists in the k-th frequency bin.

13. The computer-implemented process of claim 12 wherein the specified tracking region is defined by the expected direction of arrival of the desired signal and a maximum allowed direction of arrival error.

14. The computer-implemented process of claim 12 wherein the probability that a signal component in the k-th frequency bin originates inside a specified tracking region employs an instantaneous direction of arrival for each frequency bin based on the phase differences of the input signals.

15. The computer-implemented process of claim 12 wherein the probability that the desired signal exists in the k-th frequency bin is determined using a voice activity detector.

16. A computer-implemented process for improving the signal to noise ratio of a signal received from a microphone array, comprising:

inputting signals of sensors of a sensor array in the frequency domain defined by frequency bins and a time frame index, wherein the input signals comprise a desired signal and at least one undesired signal;

computing a beamformer output as function of the input signals divided into k frequency bins and the time frame index;

estimating the desired signal by taking the product of the beamformer output and the probability that the beamformer output contains the desired signal at a given target angle region;

using the estimated desired signal as a reference signal and subtracting components that are correlated to the estimated desired signal from the input signals to obtain interference signals;

subtracting the interference signals from the beamformer output to produce an output signal of the desired signal with an increased signal to noise ratio.

17. The computer-implemented process of claim 16 wherein at least one undesired signal has a magnitude greater than the desired signal.

18. The computer-implemented process of claim 16 wherein the input signals of the sensor array in the frequency domain are converted from the time domain into the frequency domain prior to inputting them using a Modulated Complex Lapped Transform (MCLT).

19. The computer-implemented process claim 16 wherein the probability that the beamformer output contains the desired signal at a given target angle region is computed by multiplying the probability that a signal component in a k-th frequency bin originates inside a specified tracking region by the probability that the desired signal exists in the k-th frequency bin.