Low Power Voice Trigger for Acoustic Apparatus
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Previous voice trigger systems in acoustic devices, such as microphones, consume excessive power due to constant activation of digital signal processors in search of key phrases, leading to battery depletion in mobile and wearable devices.
Innovation Solution
Implementing a low power operation mechanism where the output of the acoustic activity detector is stored in internal memory via DMA techniques, periodically activating the DSP for quick data processing and allowing it to sleep when not needed, with data transferred at high frequency only when a predetermined capacity is reached, and deactivating the DSP if the key phrase is not detected.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the DSP is constantly activated to search for key phrases, then voice trigger detection accuracy is improved, but power consumption increases
Solution Approach 1:
The acoustic activity detector performs preliminary detection of acoustic activity before the DSP is activated. This preliminary action filters out unnecessary DSP activations by only triggering DSP processing when acoustic activity is detected, thereby reducing power consumption while maintaining detection accuracy for actual voice triggers
Solution Approach 2:
The DSP is activated periodically or event-driven based on acoustic activity detection rather than continuously. This periodic activation pattern allows the system to maintain voice trigger detection capability while significantly reducing power consumption by keeping the DSP in a low-power state during intervals between activations
2Speed
If the DSP is frequently woken up to process data, then detection responsiveness is improved, but battery life deteriorates
Solution Approach 1:
Acoustic activity detection is performed in advance by a low-power detector before DSP processing is initiated. This preliminary filtering ensures that the DSP is only woken up when there is actual acoustic activity to process, maintaining responsiveness to real triggers while extending battery life by avoiding unnecessary wake-ups
Solution Approach 2:
The acoustic activity detector acts as an intermediary between the microphone input and the DSP processor. This intermediary component filters and pre-processes the audio stream, only passing relevant segments to the DSP when acoustic activity is detected, thereby reducing the frequency of DSP activations and extending battery life
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
Significantly reduces power consumption by minimizing continuous DSP activation, especially in noisy environments, allowing for efficient processing of voice triggers while maintaining low power usage.
Implementation Method 1
sound energy traverses through the port, moves the diaphragm and creates a changing potential of the back plate
Data Source
AI summary
An analog signal is received from an acoustic transducer. The analog signal is converted into digital data. A determination is made as to whether acoustic activity exists within the digital data. The digital data is stored in a temporary memory storage device and a count is maintained of an amount of digital data in the temporary memory storage device. When the count exceeds a predetermined threshold, at least some of the digital data is transmitted from the temporary memory storage device to a processor.