Abstract This paper develops a new secure control scheme (SCS) with a robust dynamic compensation for cyber-physical systems (CPSs) against malicious actuator attacks. The attack model under consideration is assumed to be a Lipschitz continuous and derivable function with respect to CPS states and an unknown bounded signal. The bounded signal is generated from any linear/nonlinear finite-L2-gain dynamical system. Different from the existing SCSs on one given injection attack pattern, (a) the new one can simultaneously resist multiple types of attacks (e.g., dynamic attacks, state-dependent attacks, etc.); (b) with the help of cooperation between two virtual dynamical systems, the presented robust compensator achieves a better compensation to attacks, such that the proposed control scheme guarantees the asymptotical stability of the attacked systems. An illustrative example verifies the effectiveness of the presented method.

中文翻译：

---

针对多种类型的执行器攻击的网络物理系统的鲁棒动态补偿方法

摘要 本文开发了一种新的安全控制方案 (SCS)，具有针对恶意执行器攻击的网络物理系统 (CPS) 的稳健动态补偿。考虑中的攻击模型假设是关于 CPS 状态和未知有界信号的 Lipschitz 连续和可推导函数。有界信号由任何线性/非线性有限 L2 增益动态系统生成。与现有 SCS 在一种给定的注入攻击模式上不同，(a) 新的 SCS 可以同时抵抗多种类型的攻击（例如，动态攻击、状态依赖攻击等）；(b) 借助两个虚拟动力系统之间的合作，所提出的鲁棒补偿器实现了对攻击的更好补偿，这样所提出的控制方案保证了被攻击系统的渐近稳定性。一个说明性示例验证了所提出方法的有效性。