Wearable audio device with feedback instability control
AssigneeBose Corporation
InventorBenjamin Jason Krosner
Life of this patent
Estimating the term…
Abstract
Aspects include approaches for feedback instability control in wearable audio devices. In certain cases, a method of controlling feedback instability in a wearable audio device with an active noise reduction (ANR) system includes: determining a current feedback instability by combining outputs from multiple instability detectors, applying latch logic to the current feedback instability to determine a current mitigation value, and adjusting a driver command signal based on the current mitigation value to mitigate feedback instability.
Claims19
A method of controlling feedback instability in a wearable audio device with an active noise reduction (ANR) system, the method comprising:
- determining a current feedback instability by combining outputs from multiple instability detectors,
- applying latch logic to the current feedback instability to determine a current mitigation value, wherein applying the latch logic comprises:
- applying a baseline mitigation value associated with a baseline instability as the current mitigation value, and
- modifying the current mitigation value in response to determining that the current feedback instability exceeds the baseline instability as controlled by a latch mechanism, wherein modifying the current mitigation value is performed by:
- comparing the current instability with the baseline instability,
- if the current instability does not exceed the baseline instability:
- controlling the baseline mitigation value based on satisfying an unlatch timer, and
- if the current instability exceeds the baseline instability:
- selecting the current instability as the current mitigation value, and
- controlling the baseline mitigation value based on satisfying a latch timer, and
- adjusting a driver command signal based on the current mitigation value to mitigate feedback instability.
A wearable audio device comprising:
- an electro-acoustic transducer for providing an audio output,
- at least one microphone configured to detect noise;
- and
- a controller coupled with the electro-acoustic transducer and the at least one microphone, the controller including an active noise reduction (ANR) system for controlling noise in the audio output, wherein the controller is configured to:
- determine a current feedback instability by combining outputs from multiple instability detectors,
- apply latch logic to the current feedback instability to determine a current mitigation value, wherein applying the latch logic includes:
- applying a baseline mitigation value associated with a baseline instability as the current mitigation value, and
- modifying the current mitigation value in response to determining that the current feedback instability exceeds the baseline instability as controlled by a latch mechanism,
- wherein the baseline mitigation value is applied to the ANR system until a triggering event is detected, wherein the triggering event includes at least one of:
- power cycling of the wearable audio device, switching operating modes of the wearable audio device, activating the ANR system, deactivating the ANR system, detecting a donning of the wearable audio device, or detecting doffing of the wearable audio device, and
- adjust a driver command signal based on the current mitigation value to mitigate feedback instability.
A method of controlling feedback instability in a wearable audio device with an active noise reduction (ANR) system, the method comprising:
- determining a current feedback instability by combining outputs from multiple instability detectors,
- applying latch logic to the current feedback instability to determine a current mitigation value,
- adjusting a driver command signal based on the current mitigation value to mitigate feedback instability, wherein adjusting the driver command signal based on the current mitigation value smooths a transition between ANR settings,
- determining an instability error value for the ANR system based on a comparison of the current instability value to a setpoint instability value, and
- applying a gain to the instability error value.
Dependencies are inferred from the claim text.