Synchronization Of Fuzzy Memristive Neural Network

With the rapid development of modern communication technology and electronic industry,daily life and scientific fields such as multimedia transmission,weather prediction,digital image processing,satellite navigation,robot control,intelligent wear and other aspects,require higher and higher information processing speed and efficiency.Therefore,the development of complex network of efficient real-time intelligent information processing has become an important research focus in the field of scientific research.The memristor neural network improves and renovates the traditional neural network by taking advantage of the automatic memory,integration convenience,nano size and hysteresis loop volt-asymmetry characteristics of memristor and adopting multi-threshold neuron properties.It reduces the power consumption and volume of complex network in production practice,and enhances the system's ability to process information efficiently and intelligently in real time.In this paper,the synchronization of fuzzy memristive neural networks is studied.The main work includes the following aspects:(1)The robust adaptive lag synchronization problem of fuzzy memristor neural networks with time-varying delays is studied.A memristor neural network model based on fuzzy model and uncertain parameters is constructed.By using Lyapunov functional method and new control rules for uncertain parameters,the sufficient and necessary conditions for robust adaptive lag synchronization of master-slave systems in memristor neural networks under the control rules are given.In addition,the robust adaptive lag synchronization of memristor neural networks in fuzzy systems is studied by combining fuzzy systems with LMI algorithm.Finally,the effectiveness and practicability of the results are verified by two simulation examples.(2)The exponential lag synchronization of memristor neural networks with reactiondiffusion terms is studied based on neural activation function control and fuzzy model.This chapter introduces a new memristor circuit which can effectively reflect the hysteresis volt-asymmetry characteristics of memristor,and carries out detailed modeling and theoretical analysis of the memristor neural network with reaction-diffusion term and the memristor neural network with fuzzy model.Furthermore,by using the Lyapunov functional method and the neural activation function controller which depends on the output of the system in the case of encapsulated circuits,theoretical conditions are obtained to ensure exponential lag synchronization of memristor neural networks with fuzzy models and reaction-diffusion terms.Finally,the correctness of the theoretical results is verified by two simulation examples.(3)The finite time synchronization problem of time-delay memristor neural networks with interval parameters based on fuzzy logic system model and nonlinear coupling is studied.This chapter constructs a model of memristor neural network based on T-S fuzzy logic system,which has great advantages in the construction of artificial brain.In order to solve the problem of unexpected parameter mismatch when selecting different initial conditions for weight matrix of delayed memrior neural network(DMNN),Filippov theory and interval matrix method are considered in this chapter to reduce the conservatism of system weight parameters and construct DMNN mathematical model more accurately and reasonably.At the same time,the nonlinear coupling in complex networks is considered in this chapter.By using Lyapunov functional method and a new discontinuous controller with discontinuous state feedback and adaptive terms,the theoretical conditions of finite time synchronization of DMNN with fuzzy system and nonlinear coupling under discontinuous state control and adaptive control are given respectively.Finally,the validity of the theoretical results is verified by two simulation examples.