Observer-based non-fragile control for robust synchronisation of switched reaction-diffusion neural networks with DoS attacks

This study intends to develop a resilient observer-based control for synchronisation of switched master-slave neural networks with reaction-diffusion terms. As well, to reflect the real-world settings, the parameter uncertainties and gain perturbations are considered in the system and control framework, respectively. Moreover, they are presumed to occur in a random manner and adhere to Bernoulli-distributed random variables. Furthermore, our work develops an observer framework for the considered network by utilising available measurement output to estimate the immeasurable states. Meanwhile, before the output signals are sent into the built-in observer, they are quantised using a logarithmic quantiser. Importantly, denial-of-service attacks are incorporated in the observer framework, thereby enhancing the resiliency of the observer. Thereafter, an attack-resilient observer-based controller is designed to accomplish the desired synchronisation. Additionally, the gain perturbations are considered to enhance the robustness of the controller. Following this, by exploiting Lyapunov stability theory, adequate criteria are established in the frame of linear matrix inequalities to ensure input–output finite time mean square synchronisation of the undertaken network model. Eventually, the simulation results are presented to demonstrate the effectiveness and utility of the theoretical findings.