Automatic Command-Detection Threshold Calibration
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing voice-activated devices face challenges in setting an initial command-detection threshold when the signal levels of spoken commands are unknown, particularly in diverse electronic environments, leading to potential misdetection of commands and user dissatisfaction.
Innovation Solution
A method that automatically sets and adjusts a detection threshold for spoken commands by analyzing ambient noise and command sounds, using noise-based and command-based thresholds, which are continually refined as more data becomes available, ensuring reliable and quick detection of commands.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If a fixed detection threshold is pre-set during device manufacture, then the device complexity is reduced and ease of manufacture is improved, but the adaptability to different electronic environments deteriorates and command detection reliability decreases
Solution Approach 1:
The system performs preliminary noise characterization during a calibration period before normal operation begins. The calibration module collects noise data during predetermined time periods when no commands are expected, establishing baseline noise levels and variance. This preliminary action enables the system to adapt to specific electronic environments without requiring complex manual configuration during manufacture.
Solution Approach 2:
The detection threshold is transformed from a static pre-set value to a dynamic parameter that automatically adjusts based on observed noise conditions. The system continuously monitors noise levels and recalibrates thresholds during operation, allowing the same device to adapt to different electronic environments and noise characteristics without requiring different hardware configurations.
2Reliability
If the detection threshold is set low to capture weak commands, then command detection sensitivity is improved, but false alarm rate increases due to noise
Solution Approach 1:
The system dynamically adjusts detection threshold parameters based on observed noise characteristics. By analyzing noise levels and variance during calibration periods, the system optimizes threshold values to distinguish between noise fluctuations and genuine commands. This parameter adaptation enables the system to maintain high sensitivity while suppressing false alarms caused by environmental noise.
Solution Approach 2:
The calibration module implements feedback mechanisms where detection performance is continuously monitored and used to adjust threshold parameters. The system learns from detection outcomes and noise patterns, refining its threshold settings over time to minimize false alarms while maintaining command detection sensitivity across varying operational conditions.
3Object-generated harmful factors
If the detection threshold is set high to reduce false alarms, then false alarm rate decreases, but command detection reliability deteriorates due to missed commands
Solution Approach 1:
The system adapts threshold parameters based on the specific electronic environment and noise characteristics rather than using conservative fixed values. By characterizing the noise profile during calibration and adjusting thresholds accordingly, the system achieves optimal detection performance for each device, ensuring commands are detected reliably without excessive false alarms.
4Measurement precision
If manual calibration is performed for each device, then detection precision is improved, but the time required for setup increases and ease of operation deteriorates
Solution Approach 1:
The calibration module enables devices to perform self-calibration automatically during operation. The system collects noise data, analyzes environmental characteristics, and adjusts detection thresholds without requiring manual intervention from users or technicians. This self-service approach maintains high detection precision while eliminating time-consuming manual calibration procedures.
Solution Approach 2:
The system performs calibration actions automatically during predetermined time periods and startup sequences, completing the precision-tuning process before normal operation begins. This preliminary automated calibration eliminates the need for manual setup while ensuring detection precision is optimized for each specific device and environment.
5Measurement precision
If a long calibration period is used to accurately characterize noise, then detection precision is improved, but the time required before normal operation increases
Solution Approach 1:
The calibration module operates periodically during predetermined time periods when no commands are expected, collecting noise data in structured intervals. This periodic approach enables accurate noise characterization to be achieved efficiently without requiring continuous or excessively long calibration periods, balancing detection precision with operational readiness.
Solution Approach 2:
The system performs essential calibration actions during startup sequences and predetermined time periods before normal operation begins, completing the most critical noise characterization quickly. This preliminary calibration approach ensures detection precision is sufficient for operation without requiring excessively long calibration periods that would delay device availability.
Data Source
AI summary
When a voice-activated device or application is first started, the signal levels corresponding to spoken commands are initially unknown, making it difficult to set detection thresholds. The inventive method provides an initial command-detection threshold based on the noise level alone. The first command is detected using this initial threshold. Then the threshold is revised according to the first command sound, and a second command is detected using the revised threshold. After detecting each command, the detection threshold is further refined according to the current noise and command sounds. Methods are also disclosed for optimizing the thresholds, adjusting parameters according to sound, and detecting voiced and unvoiced sounds separately. These capabilities enable many emerging voice-activated products and applications.