Multivalue memristor self-adaption programming circuit for nano cross rod structure and method

Abstract

The invention discloses a multivalue memristor self-adaption programming circuit for a nano cross rod structure and a method, and aims to solve a problem of precision of memristor multivalue programming and improve the efficiency of programming. The programming circuit is composed of a nano cross rod and N feedback control units; the nano cross rod comprises M horizontal nanowires, N vertical nanowires and M*N memristors; and each feedback control unit is connected with a longitudinal nanowire of the nano cross rod. The programming method comprises steps of: first conducting an erase operation on a row of memristors to be programmed; initializing resistance of the row of memristors to a high value; conducting a programming operation on the row of memristors to be programmed; each feedback control unit receiving a reference voltage signal through a reference voltage input terminal and programming the resistance values of the row of memristors to be programmed to a respective target resistance value. The resistance value of the multivalue memristor provided by the invention can be programmed and modified in a range of variation according to requirements, and the programming circuit has high precision and high efficiency of programming.

Description

technical field [0001] The invention relates to the field of nanoelectronics, in particular to a multi-value memristor self-adaptive programming circuit and a programming method oriented to a nano-cross bar structure. Background technique [0002] Memristor is a new type of nanoelectronic device, which is considered as the fourth basic electronic component alongside resistors, capacitors and inductors. Under the action of external voltage, the resistance value of memristor can change within a certain range, and keep the resistance value unchanged after power off. Based on this characteristic, in important fields such as high-density memory and reconfigurable logic circuits, a large number of research results have been proposed in the world, and it has broad application prospects. [0003] The multi-value storage property of memristor is the key to its application in high-density non-volatile memory and artificial neural network circuits and other fields. However, the appli...

AI Technical Summary

Problems solved by technology

[0005] However, the adaptive programming circuits for multi-valued memristors that have been proposed cannot meet the requirements of the nano-crossbar structure, and the programming efficiency is also low.

Existing memristor programming circuits for nano-crossbar structures are mainly based on memristors with binary storage characteristics, which cannot be directly applied to the programming of multi-valued memristors.

Currently, there is no effective solution to the programming problem of multi-valued memristors in nanocrossbar structures

Images