Abstract: A device capable of using data from multiple sensors to determine an estimated position/direction of a user with respect to the device. The device may use estimated position data, along with confidence data, that originated from a plurality of sensors to fuse the data to determine the user's estimated position and comprehensive confidence of the estimated position. The system may use the location information to perform beamforming/beam steering and/or other downstream operations using the comprehensive estimated position.

Abstract: In various embodiments, a method for estimating a room layout includes receiving candidate wall distance estimates for walls in an acoustic environment and confidence scores; receiving device location data for devices; determining, based on the device location data, a location of each wall relative to each audio processing device; generating a room layout matrix that includes the candidate wall distance estimates, wherein each row of the room layout matrix is associated with a respective device and each column is associated with a respective wall and each of the candidate wall distance estimates is associated with the respective walls based on the confidence scores; determining, based on a highest confidence score in each column of the room layout matrix and the location data, a set of wall distance estimates from an ordinal device; and determining, based on the set of wall distance estimates, room layout estimates of the acoustic environment.

Abstract: In various embodiments, a computer-implemented method comprises emitting a reference audio signal using one or more speakers of an audio processing device located in an acoustic environment, receiving, by each microphone in a microphone pair of the audio processing device, a response of the acoustic environment to the reference audio signal, determining an impulse response of the acoustic environment based on the reference audio signal and the received response of the acoustic environment, determining, based on the impulse response, a plurality of candidate wall distance estimates and a plurality of confidence scores, and determining based on the plurality of candidate wall distance estimates and confidence scores, including candidate wall distance estimates and confidence scores from one or more other audio processing devices, a set of wall distance estimates, each wall distance estimate in the set of wall distance estimates being associated with a different wall in a plurality of walls.

Abstract: Disclosed embodiments include techniques for determining spatial impulse response via acoustic scrambling. These techniques include a computer-implemented method for generating a frequency sweep signal, the method comprising generating a frequency sweep signal having a monotonically increasing frequency, partitioning the frequency sweep signal into N input segments, each of the N input segments representing a different frequency range, generating an encoding key having a sequence of N non-consecutive numbers, wherein each number in the sequence appears once, generating an output signal by selecting each of the N input segments in an order based on the sequence of N non-consecutive numbers in the encoding key, and causing a speaker to produce audio tones in an audio space based on the output signal.

Abstract: Techniques for improving sound source localization (SSL) are provided. A method for probabilistic SSL using a deep neural network (DNN) may include receiving audio data including a representation of audio such as a wakeword from a microphone array. The audio data may be processed by a DNN to output a plurality of values where each value indicates a probability that the audio originated from a direction corresponding to that value. A sensor may provide computer vision or other data which may be used to inform the plurality of values based on detecting presence of a human or obstacle. A probability that the audio originated from one of the directions of the plurality of directions may be determined based at least in part on the DNN output and the computer vision or other data.

Abstract: A device capable of autonomous motion may move in response to a user speaking an utterance, such as a command. Before moving, the device processes audio data received from a microphone array to identify different audio signals arriving at the device from different directions. Based on properties of the audio signals, the device determines which of the audio signals are merely reflections of other audio.

Abstract: A voice-controlled device includes a beamformer for determining audio data corresponding to one or more directions and a beam selector for selecting in which direction a source of target audio lies. The device determines magnitude spectrums for each beam and for each frequency bin in each beam for each frame of audio data. The device determines frame-by-frame changes in the magnitude and filters the changes to smooth them. The device selects the beam having the greatest smoothed change in magnitude as corresponding to speech.