Integrate-and-fire (I&F) neuron circuit using resistive-switching random access memory (RRAM)

We proposed the analog integrate and fire (I&F) neuron circuit using resistive-switching random access memory (RRAM). And we emulated the essential characteristics of the biological neuron such as temporal integration, threshold triggering, output generation and refractory period. The RRAM cell was made by Ag/Si₃N₄/TiN structure and measured DC I-V curve and transient pulse characteristics. We modelled the RRAM operation by spice simulation. With the RRAM model and conventional CMOS device, we simulated the I&F neuron circuit using smart spice tool. It was possible to reduce the power consumption, delay and neuron circuit cell size effectively by using the RRAM without capacitor.