Tactile Augmentation of Material Classification via Imperceptible On-Skin Triboelectricity Collection

Area of Science:

• Biomedical Engineering
• Materials Science
• Machine Learning

Background:

• Triboelectric signals from human skin offer potential for tactile perception enhancement.
• Challenges include unobtrusive signal collection, motion artifact mitigation, and managing inter-person signal variability.

Purpose of the Study:

• To demonstrate imperceptible collection and machine learning-based processing of triboelectric signals for tactile augmentation.
• To classify materials based on triboelectric signals generated during touch.

Main Methods:

• A substrate-less microfiber electrode was used on a finger to form a body-coupled triboelectric circuit.
• Triboelectric signals containing material-specific features were collected during hand-object contact.
• A machine learning model was developed for signal processing and material classification.

Main Results:

• Triboelectric signals were imperceptibly collected and processed.
• Six different materials were classified with approximately 95% prediction accuracy.
• The material differentiation model achieved approximately 88% accuracy across different users.

Conclusions:

• The body-coupled triboelectric circuit shows potential for tactile augmentation.
• Machine learning effectively processes triboelectric signals for material identification.
• This technology could lead to advanced haptic feedback systems.