State Estimation for Delayed Memristive Neural Networks With Multichannel Round-Robin Protocol and Multimodal Injection Attacks

This article explores the state estimation problem of memristive neural networks (MNNs) with mixed delays subject to multichannel round-robin protocol (MRRP) and multimodal injection attacks (MIAs). MRRP allows for multichannel-enabled signal transmission between sensors and remote estimators, thereby significantly mitigating network congestion. Different types of attacks are supposed to be encountered during the sensor data transmission via network. Both the impacts of the MRRP and the attacks are reflected in modeling the addressed system. The purpose of this article is to design an estimator capable of constraining the estimation error within an ellipsoidal region in the presence of noise interference, mixed delays, MRRP, and MIAs. Using recursive matrix inequality (RMI) techniques, sufficient conditions are derived for achieving the desired performance and the expected state estimator gains are derived by minimizing the constrained ellipsoid region. Ultimately, the efficacy of the developed estimate strategy is demonstrated through numerical simulation.