Spatially based adaptive filter for a double talk recovery method

1 . A computer-implemented method when executed by data processing hardware causes the data processing hardware to perform operations comprising:

receiving a first audio signal and a second audio signal simultaneously at an adaptive filter of an acoustic echo canceler (AEC), the first audio signal corresponding to a third-party source and the second audio signal corresponding to a target source;

estimating an echo signal present in the second audio signal based on the simultaneous first audio signal and the second audio signal received at the adaptive filter;

estimating an echo signal directivity of the estimated echo signal;

determining a target directivity of the target source;

determining a directional target-to-echo signal ratio based on the estimated echo signal directivity and the determined target directivity;

determining an adaptive floor of the directional target-to-echo signal ratio utilizing recursive averaging;

comparing the estimated echo signal directivity with the determined target directivity;

detecting that a current target-to-echo signal ratio exceeds the adaptive floor by a threshold;

detecting a divergence of an estimated echo signal direction toward a target source direction based on the detected threshold exceedance to determine AEC divergence due to a double talk scenario; and

resetting the adaptive filter of the AEC.

2 . The method of claim 1 , wherein determining the target directivity includes retrieving the target directivity from a beamformer.

3 . The method of claim 1 , wherein determining a target directivity includes utilizing a delay and sum model.

4 . The method of claim 1 , wherein determining a target directivity includes training acoustical transfer function estimations of a controller of a vehicle and utilizing geometric properties of the vehicle.

5 . A system comprising:

data processing hardware; and

memory hardware in communication with the data processing hardware, the memory hardware storing instructions that when executed on the data processing hardware cause the data processing hardware to perform operations comprising:

receiving a first audio signal and a second audio signal simultaneously at an adaptive filter of an acoustic echo canceler (AEC), the first audio signal corresponding to a third-party source and the second audio signal corresponding to a target source;

estimating an echo signal present in the second audio signal based on the simultaneous first audio signal and the second audio signal received at the adaptive filter;

estimating an echo signal directivity of the estimated echo signal;

determining a target directivity of the target source;

determining a directional target-to-echo signal ratio based on the estimated echo signal directivity and the determined target directivity;

determining an adaptive floor of the directional target-to-echo signal ratio utilizing recursive averaging;

comparing the estimated echo signal directivity with the determined target directivity;

detecting that a current target-to-echo signal ratio exceeds the adaptive floor by a threshold;

detecting a divergence of an estimated echo signal direction toward a target source direction based on the detected threshold exceedance to determine AEC divergence due to a double talk scenario; and

resetting the adaptive filter of the AEC.

6 . The system of claim 5 , wherein determining the target directivity includes retrieving the target directivity from a beamformer.

7 . The system of claim 5 , wherein determining a target directivity includes utilizing a delay and sum model.

8 . The system of claim 5 , wherein determining a target directivity includes training acoustical transfer function estimations of a controller of a vehicle and utilizing geometric properties of the vehicle.

9 . A computer-implemented method for an adaptive filter system of a vehicle, the method when executed by data processing hardware causes the data processing hardware to perform operations comprising:

receiving a first audio signal and a second audio signal simultaneously at an adaptive filter of an acoustic echo canceler (AEC), the first audio signal corresponding to a third-party source and the second audio signal corresponding to a target source;

estimating an echo signal present in the second audio signal based on the simultaneous first audio signal and the second audio signal received at the adaptive filter;

estimating an echo signal directivity of the estimated echo signal;

determining a target directivity of the target source;

determining a directional target-to-echo signal ratio based on the estimated echo signal directivity and the determined target directivity;

determining an adaptive floor of the directional target-to-echo signal ratio utilizing recursive averaging;

comparing the estimated echo signal directivity with the determined target directivity;

detecting that a current target-to-echo signal ratio exceeds the adaptive floor by a threshold;

detecting a divergence of an estimated echo signal direction toward a target source direction based on the detected threshold exceedance to determine AEC divergence due to a double talk scenario; and

executing a counter measure of the AEC.

10 . The method of claim 9 , wherein determining the target directivity includes retrieving the target directivity from a beamformer.

11 . The method of claim 9 , wherein determining a target directivity includes utilizing a delay and sum model.

12 . The method of claim 9 , wherein determining a target directivity includes training acoustical transfer function estimations of a controller of a vehicle and utilizing geometric properties of the vehicle.

13 . The method of claim 9 , wherein executing a counter measure includes resetting the adaptive filter of the AEC.