WEBINAR 6 on “Artificial Sensors: towards human-like sense of touch by means of neuromorphic approach”
Conventional electronic (e-) skin, as a large-area electronic device for sensing tactile events, comprises a sensor array, which analog signal is readout by serially sampling individual sensors in the array.
The generated frame-based data are usually utilized by machine learning based on artificial neural network (ANN) for artificial sense of touch, e.g., object classification, by touching and grasping.
The conventional e-skin has poor energy efficiency that prevents it from up-scalability, and is ill-suited for taking advantage of dynamic tactile information required for rapid tactile feedback.
In this talk, design and development of new e-skins for energy-efficient, dynamic tactile feedback are to be presented.
By means of the neuromorphic approach based on neuroscientific theory, building blocks for artificial tactile sensory systems are designed.
By means of in mixed hardware and software implementation, event-driven neuromorphic tactile systems that efficiently code dynamic tactile information for rapid object classification, and exhibit an enhancement in the spatial resolution in response to touching and grasping events are demonstrated.
WEBINAR 6 – Speaker Biography and Summary
Presentation by Dr. Zhibin Zhang and Dr. James Goodman.