Touch Modality Classification using Spiking Neural Networks and Supervised-STDP Learning

Spiking Neural Networks and synaptic learning have recently emerged as viable techniques to solve classification problems characterized by high computational efficiency when implemented on low-power neuromorphic hardware. This paper presents the implementation of a Spiking Neural Network endowed with supervised Spike Timing Dependent Plasticity for touch modality classification (e.g. poke, press, grab, squeeze, push, and rolling a wheel). Results demonstrates the ability of the network to learn appropriate connectivity patterns for the classification. The proposed network achieves a total accuracy of 88.3% overcoming similar state-of-the-art solutions.