Synchronization Control Of Stochastic Neural Network And Its Application In Chaotic Encryption And Decryption

Neural networks are complex systems composed of many neurons interconnected in a tangled manner.Actual neural networks are very complex,and a single idealized neural network model cannot fully simulate real situations.Therefore,we need to consider interfering factors of neural networks and establish more generalized and realistic neural network models.In the implementation of artificial neural networks,some physical components(such as capacitors,resistors,etc.)inherently have some uncertainty,and these uncertain factors inevitably affect the state of neural network nodes during network operation.These disturbances are usually stochastic and uncertain.Neural networks with stochastic disturbance are called stochastic neural networks,which can better simulate the state trajectory of real neural networks.In addition,the delay of information transmission and instantaneous pulse interference also affect the state of the neural network.Stochastic neural networks can exhibit very rich dynamic behaviors,among which synchronization is an important dynamic behavior.Through appropriate controllers,neural networks can achieve various types of synchronization.As a kind of chaotic system,the synchronization control of stochastic neural network is very suitable for chaotic encryption.This article explores the synchronization of stochastic neural networks and its application in chaotic encryption,and the main research contents are as follows:(1)A new predefined-time stability theorem for stochastic systems is proposed.This theorem has better flexibility and less conservatism and can be used to guide any stochastic network model to achieve predefined-time synchronization.Based on this theorem and the design of an effective controller,a new criterion for predefined-time synchronization of stochastic BAM neural networks with time-varying delays is derived using Ito’s formula.Finally,the effectiveness of the theoretical results is verified through a simulation example.(2)A new fixed-time stability theorem is proposed for stochastic systems with impulsive disturbances.Compared with the existing stability theorems,this theorem provides a more precise stability time.In order to achieve fixed-time synchronization of the system,a new feedback controller and an adaptive controller are designed.Based on this theorem and the designed controllers,fixed-time synchronization of stochastic neural networks with timevarying delays and impulsive disturbances is achieved.Finally,the effectiveness of the results obtained is verified through two simulation examples.(3)Considering more flexible control methods of impulsive systems,the predefined-time stability theorem of stochastic impulsive systems is proposed for the first time,which provides a new solution for achieving predefined-time stability of impulsive systems.Based on this theorem,a new non-chattering controller is designed to avoid the chattering phenomenon caused by the symbol function and achieve the predefined-time synchronization more quickly.The sufficient conditions for the predefined-time synchronization of stochastic impulsive neural networks are obtained.Finally,a simulation example is given to illustrate the validity of the theoretical results.(4)An encryption and decryption system based on chaotic synchronization technology for audio,images,and text is developed.By utilizing the proposed stochastic neural network synchronization control technology in this thesis,combined with the idea of chaotic encryption,the encryption and decryption functions of audio,images,and text has been realized.