The Nerve Synapse And Memory Circuit Based On New SBT Memristor

As the fourth basic electronic component,memristors are widely used in many fields,especially in synaptic bionics and resistive random access memory,because of their unique nonlinear characteristics,synaptic characteristics,nonvolatile,nano size and low power consumption.When the memristor is stimulated by the applied voltage,its memristor value will change continuously,which is similar to the regulation of the connection strength of the synaptic space of the biological nerve.In addition,memristors have the advantages of nanometer size and easy integration in morphology and structure,so they are considered to be the best way to realize artificial neural synapse.On the other hand,the memristor has the characteristics of nonvolatile,multivalued and nanoscale,which is very suitable for the design of next generation memory.In this article,a new model of SBT(Sr0.97Ba0.03 TiO3-?)memristor is established,and the working principle of SBT memristor is deeply studied.At the same time,based on SBT memristor,designed memristor synapse and memory circuit,simulated the learning behavior of synapse,and realized data storage and logic calculation.The main research contents include(1)By analyzing the measured data of SBT memristor,the corresponding mathematical model is constructed,and the basic characteristics of SBT memristor are analyzed,laying a foundation for the subsequent research on artificial synapses and storage circuits.(2)Based on the SBT memristor synaptic synaptic plasticity and higher-order learning function,the synaptic nonlinear transmission characteristics,empirical learning and long-term/short-term synaptic plasticity are realized.Furthermore,it simulates non-associative learning in accordance with human cognitive law.(3)Based on the multi-valued and non-volatile characteristics of SBT memristor,a binary memory circuit,a multi-valued memory circuit and a memristor logic circuit are designed,which can realize three kinds of logical operations:"AND","OR","XOR".The above research gives full play to the advantages of the SBT memristor,such as the nanoscale size characteristics,nonlinear characteristics,synaptic characteristics and non-volatile power failure,and proves the SBT memristor's ability to simulate artificial synapses and store information,which is ready for us to build complex artificial neural network and resistive memory in the next step.