| _version_ | 1866902031542779904 |
|---|---|
| author | VTminY |
| author_facet | VTminY |
| contents | <h1>Anti-noise TEAS-DLM</h1> <p>Deep Learning-Enhanced Anti-Noise Triboelectric Acoustic Sensor for Human-Machine Collaboration in Noisy Environments</p> <p>Main Findings Herein, an anti-noise triboelectric acoustic sensor (Anti-noise TEAS) based on a flexible nanopillars structure was developed, and was coupled with deep learning model (DLM) to identify acoustic signals for complex collaborative HMVI tasks in harsh noise scenarios. The sensor could directly acquire the acoustic fundamental frequency signals from the mixed-mode weak vibration of larynx through contact sensing with high performance. By optimizing the device layer structure, the sensor could effectively buffer the interference of the noisy environment, while the accurate meaning of the slightly distorted acoustic signals was recognized by the assistance of DLM. In the virtual world and real-world simulated scenarios, the Anti-noise TEAS-DLM was applied to transmit voice commands for guiding robots to perform complex post-disaster casualty rescue tasks in noise environment. The Anti-noise TEAS-DLM was almost 100% immune to the significant noise, efficiently guiding robots to accomplish tasks. The anti-noise capability and execution accuracy endows this DLM-coupled TEAS as a highly promising system for the collaborative HMVI tasks in complex scenarios, such as post-disaster rescue, collaborative operations, and wilderness explorations, which helps to promote the diversified and intelligent development of HMVI systems.</p> <p>The CAM_Multi_Signal.py, CAM_Single_Signal.py, Muilti_Labels_Model.py, Signal_Labels_Model.py contain the model code and the Grad-CAM code of this paper.</p> <p>Citation If you've found this work useful, please cite.</p> <p>Issues Please report issues directly to xiexi27@mail.sysu.edu.cn</p> |
| format | Recurso digital |
| id | zenodo_https___doi_org_10_5281_zenodo_15114255 |
| institution | Zenodo |
| language | |
| publishDate | 2025 |
| publisher | Zenodo |
| record_format | zenodo |
| spellingShingle | VTminY/Anti-noise-TEAS: v1.0.0 VTminY <h1>Anti-noise TEAS-DLM</h1> <p>Deep Learning-Enhanced Anti-Noise Triboelectric Acoustic Sensor for Human-Machine Collaboration in Noisy Environments</p> <p>Main Findings Herein, an anti-noise triboelectric acoustic sensor (Anti-noise TEAS) based on a flexible nanopillars structure was developed, and was coupled with deep learning model (DLM) to identify acoustic signals for complex collaborative HMVI tasks in harsh noise scenarios. The sensor could directly acquire the acoustic fundamental frequency signals from the mixed-mode weak vibration of larynx through contact sensing with high performance. By optimizing the device layer structure, the sensor could effectively buffer the interference of the noisy environment, while the accurate meaning of the slightly distorted acoustic signals was recognized by the assistance of DLM. In the virtual world and real-world simulated scenarios, the Anti-noise TEAS-DLM was applied to transmit voice commands for guiding robots to perform complex post-disaster casualty rescue tasks in noise environment. The Anti-noise TEAS-DLM was almost 100% immune to the significant noise, efficiently guiding robots to accomplish tasks. The anti-noise capability and execution accuracy endows this DLM-coupled TEAS as a highly promising system for the collaborative HMVI tasks in complex scenarios, such as post-disaster rescue, collaborative operations, and wilderness explorations, which helps to promote the diversified and intelligent development of HMVI systems.</p> <p>The CAM_Multi_Signal.py, CAM_Single_Signal.py, Muilti_Labels_Model.py, Signal_Labels_Model.py contain the model code and the Grad-CAM code of this paper.</p> <p>Citation If you've found this work useful, please cite.</p> <p>Issues Please report issues directly to xiexi27@mail.sysu.edu.cn</p> |
| title | VTminY/Anti-noise-TEAS: v1.0.0 |
| url | https://doi.org/10.5281/zenodo.15114255 |