Spike-Timing-Dependent Plasticity for a Hafnium-Oxide Memristive Synapse

Spike-timing-dependent plasticity (STDP) is a widely used technique for online learning that updates synapse weights based on the timing of pre- and post-synaptic spikes. Metal-oxide memristors, including HfO
2
, are promising in emerging synaptic circuits due to their inherent plasticity, but their non-idealities must be carefully considered in the design process. In this paper, a current-controlled memristive synapse with an STDP learning circuit is presented for improved reliability. The design limits the operating range to resistance states near the low resistance state (LRS) to account for device non-idealities at the high resistance state (HRS). This work implements a mixed-signal STDP design consisting of two parts: one programming voltage generation part, and another pulse generation part. Simulation using a 65nm CMOS process shows reliable operations both in potentiation and depression regions. Results also show good performance in terms of power and energy.