Related Experiment Video
Updated: Jul 29, 2026
12:20
Fabrication of Carbon Nanotube High-Frequency Nanoelectronic Biosensor for Sensing in High Ionic Strength Solutions
Published on: July 22, 2013
18.2K
A high recognition accuracy tactile sensor based on boron nitride nanosheets/epoxy composites for material
Shufen Wang1, Mengyu Li1, Hailing Xiang1
1School of Energy Materials and Chemical Engineering, Hefei University, Hefei City, 230601, China. dbyotw@126.com.
Materials Horizons
|April 1, 2025
Summary
This study introduces a new tactile sensor using triboelectric nanogenerators (TENGs) with advanced composite materials. This self-driven sensor accurately identifies materials, even those with negative charges, using deep machine learning.
Area of Science:
- Materials Science
- Nanotechnology
- Machine Intelligence
Background:
- Triboelectric nanogenerators (TENGs) are promising for self-driven tactile sensing in material identification.
- Existing TENG devices struggle with materials possessing electrophilic properties, leading to weak signals and poor recognition.
- There is a need for improved TENG-based sensors capable of accurately identifying negatively charged materials.
Purpose of the Study:
- To develop a TENG-based sensor utilizing boron nitride nanosheets/waterborne epoxy (BNNSs/WEP) composites for enhanced identification of negatively charged materials.
- To integrate deep machine learning with TENG technology for a robust material recognition system.
- To evaluate the sensor's performance in real-time monitoring and fatigue testing.
Main Methods:
- Fabrication of a TENG device with BNNSs/WEP composite as the friction layer.
- Characterization of the TENG device's electrical output performance (voltage and charge density).
- Development of a material recognition system incorporating deep machine learning (Convolutional Neural Network - CNN) for signal processing and material classification.
Main Results:
- The fabricated TENG device exhibited excellent output performance with a maximum voltage of 2.7 V and charge density of 88.32 nC m-2 when interacting with negatively charged objects.
- The CNN model achieved 100% accuracy in recognizing eight different materials based on TENG-generated friction electrical signals.
- The real-time monitoring sensor demonstrated high recognition accuracy for specific materials, achieving 100% for two types and varying accuracy for others.
Conclusions:
- The proposed TENG sensor with BNNSs/WEP composites significantly improves the accuracy of identifying negatively charged materials.
- The integration of deep machine learning with TENGs creates an effective system for material perception and recognition.
- This research advances the development of intelligent material identification systems within the field of machine intelligence.
