Digital implementation of a spiking neural network (SNN) capable of spike-timing-dependent plasticity (STDP) learning

The neural network model of computation has been proven to be faster and more energy-efficient than Boolean CMOS computations in numerous real-world applications. As a result, neuromorphic circuits have been garnering growing interest as the integration complexity within chips has reached several billion transistors. This article presents a digital implementation of a re-scalable spiking neural network (SNN) to demonstrate how spike timing-dependent plasticity (STDP) learning can be employed to train a virtual insect to navigate through a terrain with obstacles by processing information from the environment.