Delay Neural Network Security Event Triggered Filtering Under Dos Attack

Neural networks (NNs) are a type of artificial network system comprised of numerous interconnected basic processing units. In recent years, neural networks have been widely applied in computer technology, bioinformatics, image recognition, automation, and other fields, becoming an area of active research. In the relevant studies of neural networks, the filtering problem holds significant theoretical significance and practical value. This paper investigates the event-triggered filtering problem of delayed neural networks under Denial-of-Service (DoS) attacks. Firstly, a denial-of-service attack model is established; secondly, deep neural networks, event-triggering mechanisms, and denial-of-service attacks are integrated integrated into a cohesive framework, defining a switching filtering system and establishing a new filtering error model; then, using Lyapunov stability theory and Linear Matrix Inequality (LMI) techniques, the sufficient conditions for exponential mean-square stability of this mathematical model are derived. Reasonable boundary techniques are selected to handle delay-related terms in the Lyapunov-Krasowski functional derivative. Additionally, a new sufficient condition for the coordinated design of the filter and event-triggering parameters is derived in LMI form. At last, the proposed method's effectiveness has been validated through numerical examples.