Sleep Appliance Snore Detection Using Mobile Acoustic Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current sleep or respiratory appliances face challenges in accurately detecting snoring due to the low signal-to-noise ratio of pressure sensors, which are overwhelmed by the much larger pressure variations from the blower noise, making it difficult to detect softer snoring sounds.
Innovation Solution
Utilizing a mobile device app to record and analyze acoustic signals via a microphone, combined with sensor information from the sleep appliance, to detect snoring through acoustic models and frequency spectrum analysis, and synchronize this data with breathing cycle phases to improve snoring detection accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If pressure sensor is used for snore detection, then snoring can be detected, but the signal to noise ratio is poor due to blower noise overwhelming the snoring signal
Solution Approach 1:
The patent introduces an acoustic model as an intermediary between the pressure sensor and the snore detection process. The acoustic model processes the raw pressure signals to extract snoring characteristics, effectively filtering out blower noise and other interference. This mediator enables accurate snore detection despite the poor signal-to-noise ratio in the raw pressure sensor data.
2Measurement precision
If pressure sensor measures pressure variations, then snoring can be detected, but the full scale of the sensor is too large for detecting small pressure variations from soft snoring
Solution Approach 1:
The patent changes the parameter being measured by focusing on frequency domain characteristics rather than absolute pressure values. By transforming the pressure signal into frequency spectrum analysis, the system can detect small pressure variations from soft snoring (60 dB) that would be imperceptible in the time domain due to the large full-scale range of the pressure sensor.
3Measurement precision
If acoustic model analyzes frequency spectrum, then snoring characteristics can be recognized, but computational complexity increases
Solution Approach 1:
The patent applies partial action by focusing analysis only on specific frequency bands relevant to snoring (typically 50-150 Hz) rather than analyzing the entire frequency spectrum. This selective approach maintains high detection accuracy while reducing computational complexity compared to a full-spectrum analysis.
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 snoring detection capabilities by distinguishing snoring from background noise and breathing phases, providing improved quantitative snoring information for adjusting appliance settings, such as mask pressure and humidifier settings, thereby improving user comfort and treatment efficacy.
Implementation Method 1
record, via a microphone of the mobile device, acoustic signals during a period of use
Data Source
AI summary
A method for snore detection in a sleep appliance comprises executing an app on a mobile device and operating the sleep appliance for delivery of a pressurized flow of breathable gas to a patient interface device. The app records acoustic signals during use of the sleep appliance and analyzes, via a processor of at least one of the mobile device, the sleep appliance, or a remote cloud device, the recorded acoustic signals for generating processed audio data to detect snoring, and links, via a communication device, at least one of the recorded acoustic signals, the analyzed acoustic signals, and the recognized snoring characteristics or events between the mobile device or the remote cloud device, and the sleep appliance. Analyzing includes using an acoustic model configured to recognize snoring characteristics or events by inspecting transient acoustic pressure levels and/or a frequency spectrum in the recorded acoustic signals.