This project aims to improve current population spiking models of peripheral tactile responses such that they can be directly coupled to an appropriate touch sensor and deliver realistic tactile feedback in real-time. To this end, we will develop and validate a robust computational method that translates sensor input into spiking output, as well as modifying existing spiking models to deliver real-time output. Additionally, the touch sensors will be integrated into a setup that tracks human hand movements and contact forces, and we will collect a data set of human touch interactions during object grasping and manipulation. This data set will provide novel insights into how humans exploit their sense of touch and which tactile features underlie manipulation and grasping skills.
On the technical side, this project will deliver a method to convert touch sensor activations into realistic tactile population responses. These results will feed into neuroprosthetics and robotics, specifically through the secondment at UNIBI. On the scientific side, this project will enable a better understanding of how human use touch, and specifically which features are important for grasping and manipulation: these results will therefore guide future development of neuroprosthetic feedback algorithms and novel tactile sensors.
to study transduction in tactile afferents
to learn data analysis, feature extraction and information theory
UNIBI - SNN:
to implement neural representation of tactile stimuli
Enrolments (in Doctoral degree/s)
University of Sheffield
H. Saal, J. Wessberg, S. Panzeri, E. Chicca