Adaptive Event-Triggered Mechanism-Based Synchronization Control of Master-Slave Delayed Neural Networks Under Randomly Occurring Deception Attack

This paper investigates the problem of adaptive event-triggered mechanism-based synchronization control of master-slave delayed neural networks (MSDNNs) under randomly occurring deception attack (RODA). Firstly, in order to reduce the load of network, an adaptive event-triggered mechanism (AETM) is introduced to determine whether the sampled data should be sent to the network. By introducing a set of Bernoulli distributed random variable (BDRV), describes the probability of whether the network has been attacked and a mathematical model of RODA is established. In summary, AETM, RODA and DNNs and synchronization problems are incorporated into a unified construction, a model of the augmented system under this framework is constructed for the first time. Secondly, by employing the Lyapunov–Krasovskii functional (LKF) method and linear matrix inequalities (LMIs) techniques, some sufficient conditions on the synchronization of MSDNNs are derived, and the parameters of the designed are given in an explicit form. In conclusion, the feasibility of the method is demonstrated by a numerical example.