Practical fixed-time synchronisation of event-triggered memristor-based BAM neural networks under denial-of-service attacks

This paper investigates the practical fixed-time (PFxT) synchronisation of memristor-based bidirectional associative memory neural networks (MB-BAM-NNs) with event-triggered control (ETC) under network attacks. First, the parameter mismatch caused by the switching characteristics of memristor is addressed via convex analysis technique. Then, hyperbolic sine function based fixed-time (FxT) controller is proposed to suppress the unnecessary chattering. Also, to reduce communication load, an ETC strategy is adopted for both layers of the network during denial-of-service (DoS) attacks. Furthermore, some sufficient conditions for achieving PFxT synchronisation are established by constructing a suitable Lyapunov function and leveraging inequality techniques. Ultimately, a numerical simulation example is employed to verify the effectiveness of the theoretical analysis.