【作者】 董哲康； 闫云凤； 齐冬莲； 陈静； 段书凯；

【Author】 DONG Zhekang;YAN Yunfeng;QI Donglian;CHEN Jing;DUAN Shukai;College of Electrical Engineering,Zhejiang University;School of Electronic and Information Engineering,Southwest University;

【机构】 浙江大学电气工程学院； 西南大学电子信息工程学院；

【摘要】 自旋忆阻器是一种基于铁磁理论提出的新型二端电路元器件,其内部状态的变化取决于通过的电流密度.基于独特的存储机制以及纳米级尺寸优势,自旋忆阻器在非易失性存储、神经形态系统以及现代电路理论等诸多领域发挥了十分重要的作用.本文基于自旋忆阻器的数学模型,分析了其独特的电气性能以及阻值变化规律,得到了对应的阻值变化所需时间的求解方法.进一步,将自旋忆阻器与交叉阵列结合,提出了一种集图像存储、插值处理以及输出为一体的忆阻交叉阵列结构.一系列的计算机仿真验证了该交叉阵列结构在图像存储以及插值处理方面的有效性以及正确性,同时由于自旋忆阻器的快速阻值变化能力以及交叉阵列的并行数据处理能力,使得该方案有望促进大规模数据的快速存储与处理的一体化硬件实现.更多还原

【Abstract】 Spintronic memristor is a novel two-terminal electronic element based on the magnetic theory,whose internal state depends on the real-time current density.Due to the unique storage mechanism and nanoscale size superiority,spintronic memristor has become one of the powerful candidates in many promising applications,including nonvolatile storage,neuromorphic systems,modern circuit theory and so forth.According to the fundamental mathematical model of the spintronic memristor,this paper firstly analyzes its particular electrical property,discusses specific resistance variation law and obtains corresponding time calculation method.Meanwhile,we present a memristive structure which has the capacity for digital image storage,interpolation processing and image output by integrating the spintronic memristor into the crossbar array.The relevant effectiveness and validity of the entire scheme is verified by a series of computer simulations.Furthermore,on account of the rapid resistance variation capability and parallel processing capability,the proposed memristive system has the potential to greatly facilitate hardware implementation for large-scale data fast storage and processing.更多还原