Conditioned Reflex Mimic Circuit Design

Neural network holds great promise to be the next generation computer for the sake of its strong capacity of massive parallel processing, good capability of discerning, compatibility with analog computing, self-learning ability, and so on. In this course project, I designed a basic biological function unit cell, which can mimic one basic biological conditioned reflex function in two versions: conventional digital circuit design in Verilog HDL and the version with RRAM (Resistive Random Access Memory) as the synapse in the neuromorphic circuit. The comparison of the two versions shows that the second version with RRAM would have better performance regarding the circuit area, and thus high integrate density, non-volatile memory, bio-competitive power consumption, analog computing compatibility, and stronger circuit learning ability.