Headset Cough Detection via Multi-Transducer Fusion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional cough detectors using microphones suffer from high false acceptance rates due to incorrect identification of noises like dogs barking or people shouting as cough events, making them unreliable for diagnostics and health monitoring.
Innovation Solution
A method and apparatus for cough detection using a headset that combines signals from external and in-ear transducers, including accelerometers, to differentiate between cough events and other noises by determining energy in specific frequency bands and changes in orientation, with the option to utilize a neural network for characterizing coughs and determining associated medical conditions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional cough detectors use a single microphone to detect cough events, then the device complexity is low, but the reliability is poor due to high false acceptance rates from misidentifying other noises as coughs
Solution Approach 1:
The patent combines signals from multiple transducers including external microphones, in-ear microphones, and accelerometers to detect cough events. By merging data from these different sensing modalities, the system achieves more reliable cough detection while reducing false acceptance rates, as the combined signal analysis can distinguish coughs from other noises more effectively than a single microphone alone
Solution Approach 2:
The in-ear transducer serves multiple functions: it acts as both a microphone for acoustic signal detection and an accelerometer for motion detection. This multi-functionality allows the system to improve reliability without proportionally increasing device complexity, as one component performs multiple sensing roles
2Reliability
If conventional cough detectors monitor all frequency bands equally, then the measurement precision is simple, but the reliability is poor due to inability to differentiate coughs from other noises
Solution Approach 1:
The patent applies different signal processing techniques to different frequency bands. Specifically, it determines energy in a first frequency band using high-pass filtering and energy in a second frequency band using low-pass filtering. This localized processing approach improves reliability by capturing cough-specific frequency characteristics while managing processing complexity through targeted analysis
Solution Approach 2:
The system changes the parameter of frequency band analysis by dividing the spectrum into different bands and applying specific filtering operations to each. This parameter-based differentiation allows the system to identify cough events more reliably by examining specific frequency characteristics that distinguish coughs from other noises
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
This approach reduces false acceptance rates by utilizing both acoustic and bone-conducted signals, allowing for accurate detection and characterization of coughs, including their severity and associated medical conditions, thereby enhancing diagnostic reliability.
Implementation Method 1
receiving a second signal from an in-ear transducer of the headset
Implementation Method 2
the in-ear transducer comprises an accelerometer
Data Source
AI summary
A method of cough detection in a headset, the method comprising: receiving a first signal from an external transducer of the headset; receiving a second signal from an in-ear transducer of the headset; and detecting a cough of a user of the headset based on the first and second signals.