Fault-tolerant bumpless transfer control for fuzzy switched delayed memristive neural networks subject to false data injection attacks

This article investigates the fault-tolerant bumpless transfer (FTBT) control for fuzzy switched memristive neural networks (FSMNNs) with mixed time-varying delays and false data injection (FDI) attacks. First, the persistent dwell-time switching (PDTS) law is adopted to relax the switching frequency limitations. Next, through PDTS law and Takagi–Sugeno (T-S) fuzzy theory, the novel FSMNNs are developed considering non-ideal conditions such as mixed time-varying delays, actuator faults, and external disturbances, which improve the system reliability and align with the modeling requirement in complex situations. Meanwhile, two multiplicative variables and a stuck value are incorporated into actuator faults, capturing the dynamic evolution of fault states that are affected by practical factors. In addition, to cope with the bump phenomenon induced by switching signals and the control failures associated with actuator faults and FDI attacks, the fuzzy FTBT control scheme is employed by merely imposing bump constraints at switching instants, enhancing the flexibility of controller design. By means of Lyapunov–Krasovskii functions and inequality processing techniques, the novel conditions for the exponential stability and H∞ performance of systems are derived. Finally, the feasibility and effectiveness of the derived results are demonstrated by two simulation examples.