Performance Analysis of Spiking Neural Networks with Memristive Synapse in Detecting and Classifying RF Wave Signals.

Nowadays, the Spiking Neural Networks (SNNs) have been gaining interest because of their energy efficiency and importance in neuromorphic computing. SNN is known to exhibit good performance for processing sequential data. In this work, we conduct a simulation study to utilize SNN for classifying and detecting simple Radio Frequency (RF) wave signals, which are sequential time-series data. Our simulator implements the training procedure of SNN based on Spike Timing Dependent Plasticity (STDP) and non-linear weight update representing memristive synapses. The goal of trained SNN is to detect and classify simple RF wave signal generated with Universal Software Radio Peripheral (USRP) under various SNR conditions. Our simulation results demonstrate that SNN successfully detects the target signal in both low SNR and high SNR.