文献类型：期刊文献

英文题名：Dual-mode superhydrophobic, highly breathable proximity-tactile cellulose nonwoven sensor for speech recognition via machine learning

作者：Zhang, Rui[1,2,3];Ying, Di[1,2];Zheng, Yingying[1,2];Liu, Zhe[1,2];Wang, Jian[1,2,3,4];Zou, Zhuanyong[1,2,3]

机构：[1]Shaoxing Univ, Shaoxing Key Lab High Performance Fibers & Prod, Shaoxing 312000, Zhejiang, Peoples R China;[2]Shaoxing Univ, Natl Engn Res Ctr Fiber Based Composites, Shaoxing Sub Ctr, Shaoxing 312000, Zhejiang, Peoples R China;[3]Key Lab Clean Dyeing & Finishing Technol Zhejiang, Shaoxing 312000, Zhejiang, Peoples R China;[4]Jiangnan Univ, Sch Text Sci & Engn, Wuxi 214122, Jiangsu, Peoples R China

年份：2025

外文期刊名：CELLULOSE

收录：SCI-EXPANDED(收录号：WOS:001464388000001)、、EI(收录号：20251518222019)、Scopus(收录号：2-s2.0-105002352184)、WOS

基金：This work was supported by Shaoxing Basic Public Welfare Planning Project (2024A11019) and Zhejiang Provincial Science and Technology Innovation Program (New Young Talent Program) for College Students (2023R465032).

语种：英文

外文关键词：Carbon nanotubes; Cellulose nonwoven; Dual-mode sensor; Machine learning; Speech recognition

外文摘要：While smart textile sensors have made significant progress in the fields of health monitoring, human-computer interaction, and speech recognition, they also face many challenges, including low sensitivity, breathability, and hydrophobicity. In this study, we prepared nonwoven based sensors using ultrasound-assisted modification and dip-drying method. They have high air permeability (505 mm/s) and superhydrophobic property, with a water contact angle of 164.4 degrees, and can monitor proximity and tactile signals simultaneously. During proximity detection, the sensing distance is 13 cm with a maximum relative change of 8%, a maximum sensitivity of 3.16%/cm, and a response time of 250 ms. As far as tactile sensing performance is concerned, the sensor has a pressure sensing range of 118 kPa, high sensitivity of 1.95 kPa-1 (0-0.28 kPa), excellent cycle durability in over 2,000 pressure cycle tests, and a rapid response and recovery time (70 ms/70 ms). Additionally, the sensor is capable of detecting voice vibration signals with an accuracy of 97.5% through machine learning technology. Due to these excellent performances, it is believed that the sensor has a wide range of application prospects, including health monitoring, motion monitoring, and speech recognition.