Earbud Voice Activity Detection Using Dual-Sensor Fusion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing earbuds face challenges in accurately detecting voice activity due to noise interference and malfunctions caused by external stimuli, which can lead to incorrect determination of user utterance.
Innovation Solution
The implementation of a dual-sensor system that simultaneously uses a microphone and a bone conduction voice pick up (VPU) sensor, with dedicated filter units and voice activity detection (VAD) units, to filter and analyze signals and determine the presence of utterance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a single microphone is used for voice activity detection, then the device complexity is low, but the detection accuracy is poor due to noise interference and external stimuli
Solution Approach 1:
The patent divides the voice detection function into two separate sensor systems: a microphone for air conduction sound detection and a bone conduction VPU sensor for bone vibration detection. Each sensor has its own signal processing path with dedicated filter units and VAD units, allowing independent optimization of detection accuracy while maintaining manageable system complexity through modular architecture
Solution Approach 2:
The patent combines the outputs of two independent VAD units (one for microphone, one for VPU sensor) through a determination unit that performs logical AND operation. This merging of detection results from multiple sensors achieves higher overall detection accuracy by requiring both sensors to confirm voice activity, thereby reducing false positives from noise or external stimuli
2Measurement precision
If noise filtering is applied to improve detection accuracy, then the measurement precision improves, but the device complexity increases due to additional filter units and processing
Solution Approach 1:
The patent implements separate filter units for each sensor type: a first filter unit for microphone signals and a second filter unit for VPU sensor signals. Each filter unit is tailored to the specific characteristics of its sensor, applying appropriate noise filtering independently. This segmented approach improves signal quality without requiring a single complex universal filter, thereby managing processing complexity through specialized modular components
Solution Approach 2:
The patent applies different filtering parameters and processing methods optimized for each sensor type. The microphone path and VPU sensor path have distinct filter configurations that match their respective signal characteristics, allowing each channel to achieve optimal detection accuracy with appropriately tuned complexity rather than forcing a one-size-fits-all processing approach
3Reliability
If dual-sensor simultaneous detection is implemented, then the reliability of voice activity detection improves, but the device complexity and processing requirements increase
Solution Approach 1:
The patent creates completely separate signal processing chains for the microphone and VPU sensor, with independent filter units, VAD units, and detection logic for each sensor type. This segmentation allows each detection path to be optimized and tested independently, improving overall reliability while keeping individual module complexities manageable through modular design
Solution Approach 2:
The determination unit receives detection results from both VAD units and applies logical AND operation to determine final voice activity status. This feedback mechanism where both sensor paths must confirm detection before triggering voice activity response significantly improves reliability by eliminating false positives, while the simple logical operation maintains processing efficiency despite dual-sensor input
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This approach enables accurate detection of voice activity by comparing the detection results from both sensors, reducing false positives and improving user experience by ensuring correct interruption of content and ambient mode control.
Implementation Method 1
a second filter unit configured to filter a second signal input through a bone conduction voice pick up (VPU) sensor
Data Source
AI summary
An embodiment relates to an earbud which supports voice activity detection (VAD), the earbud comprising: a first filter unit for filtering a first signal input through a microphone; a first VAD unit for performing voice activity detection on a signal which has passed through the first filter unit; a second filter unit for filtering a second signal input through a bone conduction voice pick up (VPU) sensor; a second VAD unit for performing voice activity detection on a signal which has passed through the second filter unit; and a determination unit for comparing a detection result of the first VAD unit and a detection result of the second VAD unit to determine whether a voice activity is performed.