Adaptive Voice Detection Circuit for Speech Recognition
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing voice activity detection systems in audio devices face a trade-off between power consumption and accuracy, where high accuracy leads to complex and power-hungry voice detection algorithms, and low power consumption results in high false positive rates, causing unnecessary activation of speech recognition circuits.
Innovation Solution
A circuit and method that includes a voice detection circuit, a speech recognition circuit, and an analysis circuit to dynamically adjust input parameters based on true and false detection rates, sensitivity, and thresholds to optimize voice detection accuracy and reduce power consumption.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the voice detection algorithm is made sufficiently accurate to only activate speech recognition when voice is really present, then the false positive rate is reduced, but the voice detection algorithm becomes complex and power consumption increases
Solution Approach 1:
The patent applies dynamics by making the voice detection parameters adjustable and adaptive. The system dynamically modifies detection sensitivity, threshold values, and other parameters based on environmental conditions and operational context, allowing the algorithm to maintain high accuracy without requiring a permanently complex structure. This enables the system to optimize between accuracy and power consumption in real-time.
Solution Approach 2:
The patent implements parameter changes by modifying voice detection parameters such as sensitivity thresholds, energy computation periods, and filter characteristics. By adjusting these parameters based on environmental noise levels, usage patterns, and system state, the system achieves high detection accuracy without requiring a fixed complex algorithm structure, thereby reducing overall power consumption.
2Use of energy by moving object
If the voice detection algorithm has low power consumption, then energy is economized, but the false positive rate increases causing speech recognition to activate more often than necessary
Solution Approach 1:
The patent implements feedback mechanisms where the speech recognition circuit provides information back to the voice detection circuit about false positives and detection accuracy. This feedback loop allows the system to learn from errors and continuously optimize detection parameters, enabling low power consumption operation while maintaining high accuracy through iterative improvement rather than requiring a permanently complex algorithm.
Solution Approach 2:
The system applies self-service by enabling the voice detection algorithm to automatically adjust its own parameters based on performance metrics and environmental conditions. The algorithm monitors its own false positive rate and detection accuracy, then autonomously modifies threshold values and sensitivity settings to optimize the balance between power consumption and reliability without external intervention.
3Reliability
If speech recognition is continuously activated to ensure no voice events are missed, then detection reliability is improved, but power consumption increases significantly
Solution Approach 1:
The patent applies segmentation by dividing the speech processing system into two distinct stages: a low-power voice detection stage that continuously monitors audio input, and a high-power speech recognition stage that only activates when voice is detected. This segmentation allows the system to maintain high detection reliability through continuous monitoring while minimizing power consumption by restricting full speech recognition to only when necessary.
Solution Approach 2:
The system implements periodic action by having the voice detection circuit continuously monitor audio input at low power, and only periodically activating the speech recognition circuit when voice events are detected. This periodic activation pattern ensures no voice events are missed while avoiding the continuous power consumption that would result from keeping speech recognition constantly active.
4Reliability
If the voice detection circuit is made more sensitive to reduce false negatives, then detection accuracy improves, but false positives increase leading to unnecessary speech recognition activation
Solution Approach 1:
The patent addresses this contradiction through parameter changes by implementing multiple adjustable parameters including sensitivity thresholds, energy computation periods, zero passage thresholds, and debounce times. By independently tuning these parameters, the system can optimize the balance between reducing false negatives and controlling false positives, achieving high detection accuracy without excessive false alarm rates.
Solution Approach 2:
The system applies dynamics by making detection sensitivity and threshold parameters adaptive rather than fixed. The voice detection circuit dynamically adjusts its sensitivity and threshold values based on environmental noise levels, signal characteristics, and performance metrics, allowing it to maintain high detection accuracy while adapting to different operating conditions to minimize false positives.
Data Source
AI summary
The invention concerns a circuit for speech recognition comprising: a voice detection circuit configured to detect, based on at least one input parameter, the presence of a voice signal in an input audio signal and to generate an activation signal on each voice detection event; a speech recognition circuit configured to be activated by the activation signal and to perform speech recognition on the input audio signal, the speech recognition circuit being further configured to generate an output signal indicating, based on the speech recognition, whether each voice detection event is true or false; and an analysis circuit configured to generate, based on the output signal of the speech recognition circuit, a control signal for modifying one or more of said input parameters.