Back to results
US 10,042,038 B1Granted
Mobile devices and methods employing acoustic vector sensors
AssigneeDigimarc Corporation
InventorJohn D. Lord
Life of this patent
Estimating the term…
Abstract
Acoustic vector data is sensed via acoustic vector sensor configurations in mobile devices and used to generate sound fields. From these sound fields, positioning and orientation of the mobile device is derived. The sound fields and features derived from them are classified to provide mobile device position and other information about the environment from which the sound fields are captured. Additionally, sound fields are analyzed to detect gestural input and trigger associated programmatic actions.
Claims18
A method comprising:
- obtaining acoustic reference data associated with an environment, the acoustic reference data describing at least one attribute of a sound source from which sound can propagate into the environment;
- capturing with an acoustic vector sensor having plural axes, during a first time period, plural vector components within an acoustic field proximate to a device arranged within the environment;
- processing the captured vector components of a first acoustic vector field by beamforming to generate first acoustic vector data therefrom, the first acoustic vector data describing at least one attribute of an acoustic feature of the environment at the device;
- comparing the first acoustic vector data with the acoustic reference data;
- and
- determining a spatial arrangement of the device based on the comparing.
3 more elements
A method comprising:
- capturing acoustic vector sensor data from an acoustic vector sensor configuration having plural axes;
- converting plural axes acoustic vector sensor data into a sound map;
- and
- classifying sound sources within the sound map with a convolutional neural network to identify and locate the sound sources within the sound map, the convolutional neural network being trained to identify sound sources within the sound map by presentation of training sounds to the convolutional neural network, the training sounds being labeled with identifiers, enabling the convolutional neural network to classify sound sources within the sound map;
- wherein
- the capturing, converting and classifying is performed for plural sets of acoustic vector sensor data, each captured at different times;
- and further comprising:
- tracking location of sound sources in the sound map over the different times;
- and
- from tracked location of the sound sources, determining a spatial arrangement of the sound sources in an environment around the acoustic vector sensor configuration;
- and
- determining a location of a mobile device by matching the spatial relationship with a database of stored spatial relationships of sound sources.
9 more elements
A system, comprising:
- a housing;
- an acoustic port extending through the housing;
- an acoustic vector sensor arranged within the housing and in communication with an environment surrounding the housing via the acoustic port, the at least one acoustic vector sensor configured to capture a vector component of an acoustic field proximate to an exterior of the housing at the acoustic port and convert the captured vector component into an electrical signal;
- an output device;
- and
- a processor configured to access the electrical signal and control an operation of the output device based, at least in part, on the accessed electrical signal.
3 more elements
Dependencies are inferred from the claim text.