Microphone Array Pre-Adaptation for Low-Power Wake Word Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Battery-operated audio devices face reduced battery life due to continuous power consumption by beamformers and other complex signal processing algorithms used for detecting wakeup words and spoken commands in noisy environments, leading to degraded user experience from false positives or negatives.
Innovation Solution
Implementing pre-adaptation of beamformer filter coefficients based on environmental changes or trigger events, such as noise level, motion, or time intervals, to reduce unnecessary power usage and shorten adaptation periods.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Use of energy by stationary object
If the audio device waits for voice activity detection to trigger microphone activation, then power consumption is reduced, but response time increases and user experience deteriorates
Solution Approach 1:
The system performs preliminary actions by continuously monitoring audio levels at a lower threshold before actual voice detection is needed. This pre-conditioning allows the microphone to be activated more quickly when actual speech occurs, reducing response time while maintaining low power consumption during normal operation.
Solution Approach 2:
The system dynamically adjusts the microphone activation threshold based on detected audio levels. When audio levels exceed the first threshold, the system temporarily lowers the activation threshold to enable faster response. This dynamic adaptation resolves the contradiction by allowing quick response when needed while maintaining energy efficiency during quiet periods.
2Loss of time
If the microphone is activated immediately upon any audio detection, then response time is improved, but false activations increase and power consumption rises
Solution Approach 1:
The system segments the audio detection process into multiple stages with different thresholds. The first threshold triggers preliminary attention, while the second higher threshold confirms voice activity. This segmentation prevents false activations from immediately triggering microphone activation, improving reliability while maintaining fast response to actual speech.
Solution Approach 2:
The system dynamically adjusts the activation threshold based on the detection context. When audio levels reach the first threshold, the system enters a sensitive state where the second threshold is temporarily lowered, enabling fast response to actual voice while filtering out false triggers during normal operation.
3Reliability
If a high activation threshold is used for the microphone, then false activations are reduced, but the microphone may not activate for soft voices
Solution Approach 1:
The system dynamically adjusts the activation threshold based on the audio detection state. When audio levels reach the first threshold, the system temporarily lowers the activation threshold to the second threshold, which is optimized for detecting soft voices. This dynamic adaptation allows the system to maintain high false activation resistance during normal operation while becoming sensitive to soft voices when needed.
Solution Approach 2:
The system changes the activation parameter (threshold level) based on the detection context. By switching between two different threshold levels depending on whether the first threshold has been crossed, the system adapts its sensitivity to match the acoustic environment, preventing false activations while capturing soft voices.
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
Enhances the detection of wakeup words and spoken commands while conserving battery life by minimizing continuous power consumption and reducing adaptation time.
Implementation Method 1
the microphone is activated in response to the detected audio level
Data Source
Figure 1
Figure 2
Figure 3
AI summary
An audio device with at least one microphone adapted to receive sound from a sound field and create an output, and a processing system that is responsive to the output of the microphone. The processing system is configured to use a signal processing algorithm to detect speech in the output, detect a predefined trigger event indicating a possible change in the sound field, and modify the signal processing algorithm upon the detection of the predefined trigger event.