Objectives
Identification of neural network architectures for reproducing biological receptive field responses and implementation in neuromorphic hardware. Design, simulation, fabrication and characterization of neuromorphic circuits for spike-based computation of tactile sensory data. Evaluation of network performances with methods used in computational neuroscience to infer the information content about the stimulus (see Research Theme 6). Integration with sensors developed in Research Themes 13 and 14.
EXPECTED RESULTS
ESR15 will design and develop novel neuromorphic circuits and systems for emulating biological tactile sensing. To this end the modeling work performed in WP1 will be adapted to the analog VLSI implementation; furthermore, suitable interfaces to the sensors developed in WP3 will be designed. The final embedded system, targeted at reproducing biological RF responses, will substantially advance the state-of-the-art.