Model Construction And Features Analysis For Memristive Device With Applications

With the rapid development of modern society, more and more information needs to be stored. While the storage capacity, size and power consumption of the traditional memory devices are difficult to adapt to the speed of the information generation. The emergence of the new circuit elements-memristor and memcapacitor bring hope for problem solving. Memristive device with the characteristics of nanoscale size, automatic memory and non-linear is one of the most potential candidates of the future memory devices. Because memristor has the functions of nanoscale size and automatic memory, it has been widely applied to the cross array based on memristor as a memory unit. The cross array based on memristor has many characteristics, such as large storage capacity, low power consumption and small size, etc. When the reinforcement learning solves the complex problems, it needs a large amount of storage space. Cross array based on memristor can just reach this requirement. So the two can just combine together. Memristor with non-linear characteristic is also widely used in the chaotic system, which is based on memristor has more remarkable advantages in the size and power consumption, and the chaos caused by it can be used in secure communication, weather forecasting, image encryption and other fields.The further research of memristor and memcapacitor has been made in this thesis, their simulation and circuit models are established, their typical characteristics are also analyzed, and then the flux-controlled memristor is combined with chaotic system and reinforcement learning, meanwhile the voltage-controlled memcapacitor is applied to the single-phase bridge rectifier circuit and voltage-doubled rectifier circuit. The main content of this thesis includes the following five parts:(1) The flux-controlled memristor model is derived from the mathematical model of the HP memristor, the numerical simulations of the flux-controlled memristor are implemented, the SPICE and circuit models of flux-controlled memristor are built, and the effectiveness of the proposed model is demonstrated by a series of computer simulations.(2) Chaos and its characteristics are introduced, the complete mathematical model of the flux-controlled memristor is investigated, the flux-controlled memristor is applied to the designed chaotic system in which a chaotic attractor is observed. The dynamics characteristics of the chaotic system are analyzed. The circuit of memristor chaotic system is designed based on the mathematical model of the memristor chaotic system, meanwhile circuit simulation studies are carried out and the ideal chaotic attractor is obtained.(3) HP memristor has nanoscale size and storage characteristics, the cross array based on memristor has characteristics of large storage capacity and low power consumption. Q-learning algorithm in the reinforcement learning needs a lot of storage space while solving the more complex problems. So this thesis combines cross array based on memristor with Q-learning algorithm. The system architecture of it is designed and then it is applied in the robot obstacle avoidance experiment. The ideal results are obtained finally.(4) The Simulink and SPICE models of voltage-controlled memcapacitor are designed based on the study of voltage-controlled memcapacitor, and the simulations and verifications of these models are carried out. Meanwhile its typical characteristics are studied. The effects of the different input signals on the memcapacitor are analyzed.(5) The storage and charge-discharge characteristics of voltage-controlled memcapacitor are studied, the reading and writing circuits of voltage-controlled memcapacitor are designed. Meanwhile, compared with the general capacitance, the voltage-controlled memcapacitor has an advantage in the rectification aspect, which is applied to the single-phase bridge rectifier circuit and voltage-doubled rectifier circuit and the better rectifying effect is obtained.Finally, summaries are given and we shall explore the future prospect of the next step of the work.