US 11,539,833 B1Granted

Robust step-size control for multi-channel acoustic echo canceller

AssigneeAmazon Technologies, Inc.

InventorCarlos Renato Nakagawa

Life of this patent

Estimating the term…

Loading patent…

Abstract

A multi-channel acoustic echo cancellation (AEC) system that includes a step-size controller that dynamically determines a step-size value for each channel and each tone index on a frame-by-frame basis. The system determines that near-end signals are present by calculating a scaled error and determining that the scaled error exceeds a threshold value. When the scaled error exceeds the threshold value, the system may switch from a first cost function to a second cost function and determine a step-size value using a robust algorithm. The robust algorithm may prevent the system from diverging due to the presence of the near-end signal. For example, the robust algorithm may select a different cost function to determine the step-size value and/or combine different step-size computations, resulting in the step-size value being temporarily reduced. Thus, the robust algorithm may enable the AEC to better model the near-end disturbance statistics while the near-end signal is present.

Claims20

Dependent claims

A computer-implemented method, the method comprising:

receiving, by an autonomous motile device, a first reference audio signal;
generating, by a first loudspeaker of the autonomous motile device using the first reference audio signal, an audible sound;
receiving, from a microphone of the autonomous motile device, a first microphone signal including a first representation of the audible sound;
determining, using the first reference audio signal and a first plurality of filter coefficient values of a first adaptive filter, a first echo estimate signal that represents a portion of the first microphone signal;
determining a combined echo estimate signal by summing a plurality of echo estimate signals that includes the first echo estimate signal and a second echo estimate signal corresponding to a second reference audio signal;
determining an error signal by subtracting the combined echo estimate signal from the first microphone signal;
determining a first value by dividing an absolute value of the error signal by a second value representing a vector length of the first reference audio signal, the first value representing a scaled error for the first reference audio signal;
determining that the first value exceeds a first threshold value, the first threshold value representing a transition from a first cost function to a second cost function;
determining a first step-size value using the first reference audio signal and the error signal, wherein the first step-size value corresponds to a rate at which the first adaptive filter updates;
and
determining a second plurality of filter coefficient values for the first adaptive filter using the error signal, the first step-size value, and the first plurality of filter coefficient values.

8 more elements

A computer-implemented method performed by a device, the method comprising:

receiving a first reference signal corresponding to a first audio channel;
receiving a second reference signal corresponding to a second audio channel;
receiving a first audio input signal;
determining, using the first reference signal and a first adaptive filter, a first echo signal that represents a first portion of the first audio input signal;
determining, using the second reference signal and a second adaptive filter, a second echo signal that represents a second portion of the first audio input signal;
generating a first combined echo signal using the first echo signal and the second echo signal;
determining a first error signal using the first combined echo signal and the first audio input signal;
determining a first value using the first error signal and the first reference signal, the first value representing a first scaled error associated with the first reference signal;
determining that the first value exceeds a first threshold value, the first threshold value representing a transition from a first cost function to a second cost function;
and
determining a first step-size value using the second cost function, the first reference signal, and the first error signal, wherein the first step-size value corresponds to a rate at which the first adaptive filter updates.

8 more elements

A system comprising:

at least one processor;
and
memory including instructions operable to be executed by the at least one processor to cause the system to:
receive, by a device, a first reference signal corresponding to a first audio channel;
receive a second reference signal corresponding to a second audio channel;
receive a first audio input signal;
determine, using the first reference signal and a first adaptive filter, a first echo signal that represents a first portion of the first audio input signal;
determine, using the second reference signal and a second adaptive filter, a second echo signal that represents a second portion of the first audio input signal;
generate a first combined echo signal using the first echo signal and the second echo signal;
determine a first error signal using the first combined echo signal and the first audio input signal;
determine a first value using the first error signal and the first reference signal, the first value representing a scaled error associated with the first reference signal;
determine that the first value exceeds a first threshold value, the first threshold value representing a transition from a first cost function to a second cost function;
and
determine a first step-size value using the second cost function, the first reference signal, and the first error signal, wherein the first step-size value corresponds to a rate at which the first adaptive filter updates.

11 more elements

Dependencies are inferred from the claim text.