Texture Discrimination with a Soft Biomimetic Finger Using a Flexible Neuromorphic Tactile Sensor Array That Provides...

Area of Science:

• Biomimetic robotics
• Neuroprosthetics
• Tactile sensing

Background:

• Soft robotic fingers offer safe object manipulation.
• Prosthetic tactile sensing can improve amputee experience.
• Neuromorphic sensors mimic biological sensory processing.

Purpose of the Study:

• To integrate a soft biomimetic finger with a neuromorphic tactile sensor array.
• To enable texture discrimination using spike-based features.
• To provide sensory feedback for a more natural prosthesis.

Main Methods:

• A pneumatically actuated soft finger was combined with a textile neuromorphic tactile sensor.
• Tactile sensor outputs were converted into neuromorphic spike trains.
• Spike-based features were used with a support vector machine classifier for texture differentiation.

Main Results:

• The system achieved 99.57% average classification accuracy on 13 textured surfaces.
• Performance was evaluated across 16 parameters (finger flexion angles and palpation speeds).
• Transcutaneous electrical nerve stimulation successfully conveyed texture information to subjects.

Conclusions:

• This soft biomimetic finger with neuromorphic encoding demonstrates high accuracy in texture discrimination.
• Neuromorphic techniques and sensory feedback show potential for creating more human-like prostheses.
• Texture feedback can significantly enhance user interaction with their environment.