Application And Research Of The Memristor Crossbar Array In Logic Gates And Image Recognition

With the rapid development of computer technology, the logic circuit and image recognition have been widely used in many fields. Higher performances of these circuits are required to meet the development. Memristor, as a new generation of resistive memory device with the advantages of fast read and write speeds, low power consumption, non-volatile memory, small size, can change the methods of information storage and transformation, improve the effectiveness of computer, even change the computer architecture. Benefit from the structural characteristics of a memristor, it has a strong compatibility with CMOS technology, which makes it a very attractive prospect in the massive memory. As a storage structure, memristor, which is capable of performing both logic and memory operations, can implement logic operations directly in nano structures. Benefiting from this character, the technology of memristor will play a positive role in solving the problems of image recognition, and it is also worthy of depth discussion and research.This thesis first outlines the development and research status of memristor, then makes a brief introduction for the mechanism of memristor, analyzes several practical models and their working principle and characteristics. Anti-series memristor switch CRS is also provided. Based on I-V curves and the simulation of CRS, the relevant characteristics of CRS are described. Then the crossbar array drain current problems and solutions are analyzed. Then some logic gate circuits which are based on just memristors are designed. Using CRSs as input units, memristors as output units which stores the final result, AND, OR, NAND, NOR, NOT logic operations can be realized. This thesis also describes the development of image recognition, based on the switching characteristics of memristor and CRS, a memristor crossbar array structure for image recognition is designed. Using this structure, the differences of input data and storaged data can be seen through the change process of the capacity's electric quantity. Therefore, identification of two images can be implemented. What's more, two images are encoded through PPED algorithm and the comparison result is provided. Therefore, the correctness and authenticity of the subject are verified. The results of this research is based on non-volatile storage properties of memristor, this thesis not only complete the applications of memristor in several hot areas, but also provide a reference in other application areas.