Synchronization of Delayed Neural Networks via Integral-Based Event-Triggered Scheme

Area of Science:

• Control Systems Engineering
• Computational Neuroscience
• Information Security

Background:

• Delayed neural networks (NNs) present challenges in achieving synchronized states due to inherent time delays.
• Traditional event-triggered schemes (ETS) can be inefficient, leading to unnecessary data transmission and computation.

Purpose of the Study:

• To develop a novel integral-based event-triggered scheme (IETS) for achieving event-triggered synchronization in delayed neural networks.
• To establish sufficient conditions for controller existence ensuring asymptotic synchronization using Bessel-Legendre inequalities and linear matrix inequalities (LMIs).
• To demonstrate the superiority of the proposed IETS over existing ETS methods and apply it to image encryption.

Main Methods:

• Proposal of an integral-based event-triggered scheme (IETS) utilizing system states and past triggered data.
• Transformation of the synchronization problem into a distributed delay problem.
• Application of Bessel-Legendre inequalities to derive linear matrix inequalities (LMIs) for controller design.

Main Results:

• Sufficient conditions for asymptotic synchronization of delayed neural networks were derived using LMIs.
• Illustrative examples confirmed the enhanced performance of the proposed IETS compared to conventional ETS.
• The IETS method was successfully applied to image encryption and decryption, introducing a novel algorithm for improved quality.

Conclusions:

• The proposed integral-based event-triggered scheme (IETS) offers an effective approach for synchronizing delayed neural networks.
• The IETS method provides a more efficient control strategy than traditional event-triggered schemes.
• The developed IETS is applicable to practical problems like secure image encryption, enhancing process quality.