AI enable wireless sensing for remote speech recognition

Chen, Zikang (2025) AI enable wireless sensing for remote speech recognition. MSc(R) thesis, University of Glasgow.

Abstract

Contactless health monitoring is becoming an area of significant attention, especially after the impact of the COVID-19 pandemic. With the development of RF sensing technology, its application prospects in healthcare have garnered significant attention. Radio Frequency (RF) sensing techniques such as ultra-wideband (UWB) radar and Frequency Modulated Continuous Wave (FMCW) radar are used in many contactless monitoring scenarios. Compared to contact-based health monitoring methods, RF sensing technology offers users a non-intrusive experience, which enhances patients’ quality of life. Additionally, when compared to traditional non-contact monitoring technologies like imaging, RF sensing provides superior privacy protection, which effectively addresses users’ concerns. Artificial intelligence technology is also advancing rapidly and has gained significant attention due to its outstanding performance across various application scenarios. The integration of artificial intelligence with RF sensing technology can offer excellent and convenient solutions for future healthcare. This thesis proposed a multimodal speech recognition system of UWB radar data, acoustic information and visual information. The proposed multimodal approach achieves 96.89% accuracy in the word classification task, which indicates the performance improvement of the incorporation of the UWB for the multimodal system compared to the single-modal system.

• Item Type: Thesis (MSc(R))
• Qualification Level: Masters
• Subjects: T Technology > T Technology (General); T Technology > TA Engineering (General). Civil engineering (General)
• Colleges/Schools: College of Science and Engineering > School of Engineering
• Supervisor's Name: Abbasi, Professor Qammer
• Date of Award: 2025
• Depositing User: Theses Team
• Unique ID: glathesis:2025-85175
• Copyright: Copyright of this thesis is held by the author.
• Date Deposited: 11 Jun 2025 10:08
• Last Modified: 11 Jun 2025 10:11
• Thesis DOI: 10.5525/gla.thesis.85175
• URI: https://theses.gla.ac.uk/id/eprint/85175

Related URLs

• Conference proceeding
• Article DOI
• Article DOI