A Flexible Sponge-based Vibration Sensor for Conversion of Neck-skin Vibration to Digital Voice

Project: Research

View graph of relations


In recent years, flexible “smart skin” vibration sensors conformable to human necks have been examined as wearable voice recognition devices owing to their potential applications in various important areas such asreal-time vocal-load monitoring, novel microphones(with background noise reduction),voice-based biometrics, and voice-based disease diagnostics.However, the development of skin-conformable vibration sensors with high sensitivity, low damping, and frequency response optimized for collecting neck-skin vibration signals generated by the vocal cords, remains to be realized. In addition, similar to the development of “epidermal electronics”, system integration with wireless and self-powered components, and the identification of materials to prepare soft sensors and electronics that are compatible with body curvilinear surfaces and durable/stable against bodily activities remain significant scientific and engineering challenges. In this project, our interdisciplinary team will develop 1)an integrated wireless conformal neck-skin vibration sensor based on thin sponge structures embedded with carbon nanotubes(CNTs) to detect skin vibration, which has been shown to be highly correlated with the voice pressure. Subsequently, 2)neck-skin vibration data from volunteers will be collectedusing the skin-conforming sensors, and 3)AI algorithms will be implemented to correlate or “translate” the skin vibration signals into digital speech voice.In addition, we will 4)perform theoretical analyses and modeling to optimize the mechano-electrical responsivity and frequency response of the novel flexible CNT-sponge-based sensorsto maximize their sensitivity. According to our preliminary results, the skin-like CNT-based sponge sensor can produce piezoresistive and capacitive response signals due to input vibrations and is expected to detect human throat skin vibrations (up to ~4 kHz) and muscle movements (typically less than 10 Hz). To achieve these objectives, we will tackle several critical technological problems, specifically those related to thea) development of thin, durable, stable, and high-sensitivity sensing elements compatible with curvilinear body surfaces; b) establishment of a scalable fabrication process for thin CNT-sponge sensing devices optimized for neck-skin vibration monitoring;andc) development of AI-based algorithms to effectively correlate neck-skin vibrations to spoken words or “translate” skin vibrations into synthetic speech. The results of this project will establish Hong Kong at the forefront in global R&D development for smart sensors that can be applied in artificial larynx frameworks, voice-based healthcare, and disease diagnostics. The success of this project will help expand Hong Kong’s technological strength in healthcare-sensor development and promote the social and economic development of Hong Kong, and eventually, the world. 


Project number9043327
Grant typeGRF
Effective start/end date1/01/23 → …