Projective synchronization in fixed/predefined-time for quaternion-valued BAM neural networks under event-triggered aperiodic intermittent control.

This study aims to solve the fixed/predefined-time projective synchronization (FTPS/PTPS) of quaternion-valued BAM neural networks (BAMNNs) through event-triggered aperiodic intermittent control (ETAIC). Firstly, a novel quaternion-valued BAMNN model is established by integrating discontinuous activations, parameter uncertainties and time-varying delays. Subsequently, under the framework of the non-separation method several concise ETAIC schemes are designed by utilizing an exponential term instead of a sign function and two power-law terms. Different from the previous time-triggered intermittent controllers, the control and rest intervals of the ETAIC strategies are determined according to the dynamics of the network rather than being pre-specified. Furthermore, several adequate conditions are acquired to guarantee the FTPS and PTPS of the considered networks by constructing auxiliary functions and applying Taylor expansion Ultimately, three simulation examples are presented to confirm the validity and usefulness of the theoretical findings. • A novel quaternion-valued BAM neural network is established. • Several concise event-triggered aperiodic intermittent control schemes are designed. • New sufficient criteria are obtained. [ABSTRACT FROM AUTHOR]