基于动态输出反馈的区间二型T-S模糊系统的模型预测安全控制  

Model Predictive Security Control of Interval Type-2 T-S Fuzzy System Based on Dynamic Output Feedback

作　　者：马江涛 宋燕[2] 

机构地区：[1]上海理工大学理学院,上海 [2]上海理工大学光电信息与计算机工程学院,上海

出　　处：《建模与仿真》2023年第3期1908-1925,共18页Modeling and Simulation

摘　　要：本文研究了一类在测量输出端带有欺骗攻击的区间二型T-S模糊系统模糊模型预测控制(FMPC)的安全问题。针对不可测的系统状态、系统的非线性和欺骗攻击的破坏性,采用了FMPC框架下的动态输出反馈控制,同时提出了无穷时域上的最坏情况的优化问题,进行性能分析和控制综合。利用二次函数方法和奇异值分解技术,解决了变量之间耦合引起的非凸性,并导出了满足终端约束集的条件。此外,为了减轻攻击对递推可行性的破坏,引入了特殊标量,应用了不等式分析技术。在此基础上提出了一个具有可解性的辅助优化问题来求出所需的控制器,并得到了保证在基于FMPC的控制器下控制系统在H2意义上是均方安全的充分条件。最后,通过一个算例证明了所提方法的有效性。This paper studies the security of fuzzy model predictive control (FMPC) for a class of interval type-2 T-S fuzzy systems with deception attack at the measurement output. Aiming at the unmeasurable system state, the nonlinearity of the system and the destructiveness of the deception attack, the dynamic output feedback control under the FMPC framework is adopted, and the optimization problem of the worst case in the infinite time domain is proposed for performance analysis and control synthesis. By using the quadratic function method and singular value decomposition tech-nique, the non-convexity caused by the coupling between variables is solved, and the conditions to satisfy the terminal constraint set are derived. In addition, in order to reduce the damage of attack to the recursion feasibility, a special scalar is introduced and inequality analysis technology is ap-plied. On this basis, a solvable auxiliary optimization problem is proposed to find the required con-troller, and a sufficient condition is obtained to ensure that the control system is mean square safe in the sense of H2 under the FMPC-based controller. Finally, an example is given to demonstrate the effectiveness of the proposed method.

关 键 词：动态输出反馈 安全控制 奇异值分解 欺骗攻击 控制系统 二次函数 模糊模型预测 可解性 

分 类 号：TP3[自动化与计算机技术—计算机科学与技术]