Dynamic Wakeword Detection Threshold Adjustment
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Speech recognition systems often fail to detect wakewords accurately when users speak multiple commands in quick succession, due to a single, consistent wakeword detection sensitivity that may not accommodate rapid utterances.
Innovation Solution
Implementing multiple wakeword detection sensitivities, allowing a device to use a lower sensitivity after detecting an initial wakeword, enabling more sensitive detection of subsequent wakewords spoken in a short time period, and adjusting sensitivity based on historical data and user-specific patterns.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a single consistent wakeword detection sensitivity is used, then the system maintains stable operation, but it fails to accurately detect wakewords in rapid sequences
Solution Approach 1:
The patent implements dynamic wakeword detection sensitivity that automatically adjusts based on the temporal patterns of detected wakewords. When multiple wakewords are detected in rapid succession, the system increases sensitivity to capture follow-up commands. This dynamic adjustment resolves the contradiction by making the detection system adaptable to different usage scenarios while maintaining overall reliability through controlled sensitivity changes.
Solution Approach 2:
The system changes the detection sensitivity parameter in response to detected usage patterns. By monitoring the time intervals between wakeword detections and adjusting the sensitivity threshold accordingly, the system can optimize detection accuracy for rapid utterances without sacrificing stable operation during normal conditions.
2Reliability
If sensitivity is increased to detect rapid wakeword sequences, then detection accuracy improves, but false positives increase
Solution Approach 1:
The sensitivity adjustment is dynamic and context-dependent rather than permanently increased. The system temporarily increases sensitivity only when rapid wakeword patterns are detected, and automatically returns to baseline sensitivity after a predetermined time period. This dynamic approach allows the system to capture rapid utterances while minimizing false positives during normal operation.
Solution Approach 2:
The system implements periodic sensitivity adjustments based on detected usage patterns. After detecting multiple wakewords in rapid succession, the system enters a high-sensitivity mode for a limited duration, then periodically returns to normal sensitivity. This periodic action pattern allows accurate detection of rapid sequences while preventing sustained high sensitivity that would generate excessive false positives.
3Adaptability or versatility
If multiple sensitivity levels are implemented, then responsiveness to rapid utterances improves, but system complexity increases
Solution Approach 1:
Rather than implementing multiple separate detection systems, the patent uses a single detection system with dynamically adjustable sensitivity. The system monitors wakeword detection patterns and automatically adjusts its sensitivity parameter, providing adaptability to rapid utterances without requiring multiple complex detection pathways. This dynamic approach achieves versatility while maintaining relatively simple system architecture.
Data Source
AI summary
Techniques for using a dynamic wakeword detection threshold are described. A device detects a wakeword in audio data using a first wakeword detection threshold value. Thereafter, the device receives audio including speech. If the device receives the audio within a predetermined duration of time after detecting the previous wakeword, the device attempts to detect a wakeword in second audio data, corresponding to the audio including the speech, using a second, lower wakeword detection threshold value.