The Study And Application Of Memristive Bridge Synapse Circuit

With the rapid development of the information technology, more and more information need to be processed. So the neuromorphic system, which has the advantage of parallel computing, is becoming more and more important in information processing. However, the development of neuromorphic system meets obstacles because the size of transistor has approached almost at the minimum. In neuromorphic system, the tradition artificial neural synapse cannot satisfy the requirement of dealing with real time and complex situation. Memristor, a new nonlinear circuit element, has properties of memory and the similar synapse characteristic. Based on these properties, Memristor is suitable to be used as the electronic synapse. Using memristor devices as synapse has been presented in many papers. Spintronic memristor is a new memristor model, which can be used to construct a simple and compact memristor bridge synapse circuit. This synaptic model is able to modify signed synaptic weighting and has the good matching features. These properties can be illustrated by simulation experiments of neuromorphic system. So the memristor synapse circuit has promising application in many fields such as very-large-scale-integrations (VLSI), image processing, artificial intelligence and so forth.In this thesis, we firstly introduce the basic theory and properties of HP-memristor and spintronic memristor, analyze their typical characteristics and establish their model simulations. Then, the voltage-controlled spintronic memristor and resistance are applied in the synapse circuit. And we discuss the advantage of classical bridge synapse circuit and spintronic memristor bridge synapse. Furthermore, the spintronic memristor bridge synapse is used in Memrisitive cellular automata. The behavor of memristive cellular automata can response to point-wise under the rule of Brian’s Brain. Considering the initial sensitivity and the complex of synapse station of the cellular automata, we design a new encryption algorithm for image encryption. Finally, the memristive bridge synapse based-on STDP is designed. At the same time, we construct the different weight template and enforce it on the application of image processing. The research of this thesis will promote the development of neuromorphic system because the simple structure of memristive bridge synapse circuit combine the precise of adjusted memristive synapse weight with the special properties of VLSI of cellular neural network. It is potential to design a new extra-high integration and bionic intelligent neuromorphic system by means of our results in the future.