Related Experiment Video
Updated: Aug 27, 2025
07:34
A Simple Stimulatory Device for Evoking Point-like Tactile Stimuli: A Searchlight for LFP to Spike Transitions
Published on: March 25, 2014
10.0K
Neuromorphic Tactile Edge Orientation Classification in an Unsupervised Spiking Neural Network
Fraser L A Macdonald1,2, Nathan F Lepora1,2, Jörg Conradt3
1Department of Engineering Mathematics, University of Bristol, Bristol BS8 1TW, UK.
Sensors (Basel, Switzerland)
|September 23, 2022
Summary
This study introduces a novel neuromorphic tactile sensor (NeuroTac) for robotic hands. It accurately detects edge orientations using an event-based optical sensor and spiking neural networks, advancing artificial touch capabilities.
Area of Science:
- Robotics
- Neuroscience
- Sensor Technology
Background:
- Dexterous manipulation in robotic systems necessitates sophisticated artificial touch.
- Current tactile sensors often lack the bio-inspired processing capabilities for complex tasks.
- Neuromorphic computing offers a promising avenue for advanced sensory processing.
Purpose of the Study:
- To investigate neuromorphic tactile sensation for edge orientation detection.
- To integrate an event-based optical tactile sensor with spiking neural networks.
- To develop a bio-inspired artificial fingertip for enhanced robotic manipulation.
Main Methods:
- An event-based vision system (mini-eDVS) was incorporated into a low-form factor artificial fingertip (NeuroTac).
- Tactile data was processed using a Spiking Neural Network with unsupervised Spike-Timing-Dependent Plasticity (STDP) learning.
- A 3-nearest neighbours classifier was employed for edge orientation classification.
Main Results:
- The NeuroTac sensor reliably detected edge orientations in 10-degree increments.
- Accurate classification was achieved during both vertical tapping and horizontal sliding motions.
- The system demonstrated effective bio-inspired tactile processing.
Conclusions:
- The developed neuromorphic tactile sensor shows high reliability in edge orientation detection.
- This technology has the potential to significantly improve tactile sensing in robotics and prosthetics.
- Further development could lead to more dexterous and adaptable artificial manipulation systems.
