Silent Speech Wearable Control With Gaze-Directed Audio Selection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing speech input technologies for portable and wearable devices are limited in noisy environments and pose privacy concerns, and current methods for controlling devices without audible input, such as hand or eye gestures, are straining or socially awkward.
Innovation Solution
Utilizing silent speech and sound inputs, like whispering or soft tongue clicks, combined with innovative microphones and eye tracking, to control devices in noisy environments while maintaining privacy and comfort.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If audible speech input is used for device control, then speech recognition accuracy is improved, but privacy is compromised and usability in crowded places deteriorates
Solution Approach 1:
The patent replaces traditional acoustic microphones with a laser-based detection system. A laser beam is directed at the user's throat area, and vibrations of the throat tissues during silent speech are detected through laser Doppler vibrometry. This optical/mechanical detection method substitutes the acoustic field-based microphone approach, enabling detection of silent speech vibrations without requiring audible sound waves, thus maintaining privacy while enabling speech recognition.
2Object-affected harmful factors
If hand or eye gestures are used for device control, then privacy is maintained, but ease of operation deteriorates due to straining or social awkwardness
Solution Approach 1:
The patent replaces manual or ocular mechanical gesture systems with a physiological vibration-based detection system. Instead of requiring users to perform deliberate hand movements or eye gestures, the system detects natural throat vibrations that occur during silent speech. This substitution eliminates the need for strained or socially awkward gestures while maintaining privacy, as the detection occurs through laser vibrometry of internal throat tissue movements rather than external acoustic signals.
3Adaptability or versatility
If speech input is used in noisy environments, then communication capability is improved, but reliability of speech recognition deteriorates
Solution Approach 1:
The patent replaces acoustic field-based speech detection with direct mechanical vibration detection of throat tissues. By using laser Doppler vibrometry to detect the physical vibrations of the throat during speech production, the system bypasses the noisy acoustic environment entirely. The laser detects vibrations at the source (throat tissues) rather than picking up sound waves that have propagated through the noisy air, thus maintaining high reliability in noisy environments while preserving communication capability.
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
Enables secure and comfortable device control in noisy conditions, enhancing speech recognition and privacy, and allowing secure entry of sensitive information without straining or social awkwardness.
Implementation Method 1
a laser is directed at a throat of the user and vibrations of a throat of the user during the silent speech are detected
Data Source
AI summary
An apparatus is described herein related to detecting and processing silent speech. The apparatus comprises a head-mountable device and at least one ear-mountable device, wherein the ear-mountable devices are configured to selectively pick up both head-internal and head-externals sounds, thus enabling silent speech to be detectable and wherein the head-mountable device comprises an eye tracking unit capable of determining a gaze dwell of a user of the device, thus enabling silent speech commands to be directed at objects subject to a gaze dwell, or enabling speech or audio originating from an interlocutor who is subject to a gaze dwell to be selectively enhanced. Methods are described herein relate to the application of semantic intent derived from a silent speech to real-word objects which are subject to a gaze dwell by a user.


