Self-Powered Intelligent Human-Machine Interaction for Handwriting Recognition.

Area of Science:

• Materials Science
• Electrical Engineering
• Computer Science

Background:

• Handwritten signatures are ubiquitous, posing challenges for effective signal recognition.
• Triboelectric nanogenerators (TENGs) offer a novel method for detecting mechanical signals, enabling active sensor development.
• Developing efficient approaches to capture and interpret handwritten signals is crucial for human-machine interaction.

Purpose of the Study:

• To develop an intelligent human-machine interaction interface utilizing a triboelectric nanogenerator.
• To enable the recording of handwritten triboelectric signals without external power.
• To achieve recognition of handwritten characters and numerals using machine learning.

Main Methods:

• A horizontal-vertical symmetrical electrode array was employed to capture triboelectric signals from handwriting.
• Supervised machine learning algorithms were combined with principal component analysis (PCA) for signal processing and recognition.
• PCA was used to reduce data complexity for neural network processing.

Main Results:

• The developed interface successfully recorded handwritten triboelectric signals.
• Handwritten English letters, Chinese characters, and Arabic numerals were accurately recognized.
• The system demonstrated potential for anticounterfeiting by analyzing writing habits.

Conclusions:

• The intelligent interface based on TENGs provides a novel solution for handwritten signal recognition.
• This technology has significant potential applications in signature security and advanced human-computer interaction.
• The combination of TENGs and machine learning offers a promising direction for future interactive systems.