Analysis and Optimization of Secure Sliding Mode Observer-Based Control in Nonlinear Descriptor Systems Under Attacks
针对受传感器和执行器攻击的非线性描述符系统，提出一种结合T-S模糊模型、Q学习事件触发机制和滑模控制的弹性安全框架，并通过秘书鸟优化算法优化控制器和观测器增益，在卡车-拖车系统仿真中验证了有效性。
This article proposes a resilient control framework for securing cyber-physical systems (CPSs), specifically addressing nonlinear descriptor systems operating under communication constraints and subject to sensor and actuator attacks. We integrate Takagi-Sugeno (T-S) fuzzy models with a Q-learning-based event-triggered mechanism (ETM) and adopt a sliding-mode control strategy to establish a resilient security architecture that adaptively balances operational efficiency with robust protection against cyber-physical threats. A major contribution of this work lies in designing an adaptive fuzzy sliding-mode observer (SMO) with mismatched premise variables for the estimation of compromised system states. Additionally, a sliding-mode controller (SMC) is synthesized to maintain closed-loop admissibility and ensure the reachability of sliding surfaces. We advance beyond the existing approaches by employing the secretary bird optimization algorithm (SBOA) to optimize controller and observer gains, thereby solving the nonconvex optimization challenges present in controller and observer design. The effectiveness of the proposed method is validated through extensive Monte Carlo simulations on a truck-trailer system. These simulations demonstrate the efficacy of the approach in maintaining system stability and performance under various attack scenarios, thereby making a significant contribution to the security of nonlinear systems in networked environments.