P‐12.8: SLTS‐ResNet: A Self‐Powered Silent Speech Recognition System Based on Laryngeal Vibration Triboelectric Sensor and Deep Learning for Assistive Communication
SID Symposium Digest of Technical Papers · 2026-08 · 57 (S1) · 1818-1820
DOI: 10.1002/sdtp.70003_485 ↗Kurz gefasst
Fachartikel in der Zeitschrift SID Symposium Digest of Technical Papers, online erschienen am 11.09.2026. Verfasst von Xinyue Yang und 5 weiteren Autor:innen. Die Arbeit nennt 5 Literaturangaben.
Abstract (Originalsprache)
This study develops a complete silent speech recognition system for assistive communication, integrating a self‐powered laryngeal vibration triboelectric sensor (SLTS) with a deep learning model based on ResNet‐SE (Residual Neural Network with Squeeze‐and‐Excitation module). The SLTS consists of a PDMS/BST composite film and silver fiber conductive fabric, forming a contact‐separation TENG that directly converts throat vibrations into electrical signals without external power. The system achieves remarkable recognition accuracy of 97.63% for 18 common phrases using a dataset of 1,440 samples collected from healthy speakers under silent articulation conditions. Furthermore, it demonstrates exceptional environmental robustness, maintaining stable signal acquisition in noisy environments (77.2 dB) and under wind interference (83.2 dB). By leveraging the individual uniqueness of laryngeal vibration patterns, the system also accomplishes user identity recognition with 97.99% accuracy across five users. This research provides a practical, wearable solution for speech‐impaired individuals, effectively bridging communication gaps while offering new technological paradigms for intelligent healthcare and biometric authentication.
Zitieren
APA
Yang, X., Li, Z., Yin, Y., Zhang, Y., Zhang, Y., & Lin, J. (2026). P‐12.8: SLTS‐ResNet: A Self‐Powered Silent Speech Recognition System Based on Laryngeal Vibration Triboelectric Sensor and Deep Learning for Assistive Communication. SID Symposium Digest of Technical Papers, 57(S1), 1818-1820. https://doi.org/10.1002/sdtp.70003_485
Vancouver
Yang X, Li Z, Yin Y, Zhang Y, Zhang Y, Lin J. P‐12.8: SLTS‐ResNet: A Self‐Powered Silent Speech Recognition System Based on Laryngeal Vibration Triboelectric Sensor and Deep Learning for Assistive Communication. SID Symposium Digest of Technical Papers. 2026;57(S1):1818-1820. doi:10.1002/sdtp.70003_485
BibTeX
@article{Yang2026,
author = {Yang, Xinyue and Li, Ziliang and Yin, Yuwen and Zhang, Yu and Zhang, Yongai and Lin, Jianpu},
title = {P‐12.8: SLTS‐ResNet: A Self‐Powered Silent Speech Recognition System Based on Laryngeal Vibration Triboelectric Sensor and Deep Learning for Assistive Communication},
journal = {SID Symposium Digest of Technical Papers},
year = {2026},
volume = {57},
number = {S1},
pages = {1818--1820},
doi = {10.1002/sdtp.70003_485}
}
EUVIMED