Exponential Synchronization of Quaternion-Valued Delayed Neural Networks under Sampled-Data Control

The article conducts the research on exponential synchronization of quaternion-valued neural networks (QVNNs). In order to save the consumption of bandwidth resources, a sampled-data mechanism is employed to achieve synchronization. At first, a distinctive two-sided looped Lyapunov functional (LF) is established with some new augmented items, which makes full use of feasible and effective sampling information and system characteristics. The difficulty caused by the complex model is decreased by means of the real decomposition method and transforming the error system to an augmented system. Then, some conditions are derived to synchronize the quaternion-valued delayed neural networks (QVDNNs) in this article which can generalized the existing results. Furthermore, by leveraging linear matrix inequalities (LMIs) method, the desired gain matrix can be obtained. Lastly, the feasibility of the derived results is further explained through simulation.