Optimal Stealthy Attack Design Using Sequential Linear Programming

This article presents a novel methodology based on Sequential Linear Programming (SLP) for designing optimal (stealthy) attack sequences aimed at maximizing a nonlinear attack function. Our investigation extends to scenarios where adversaries can inject attack signals into both control in-put and sensor measurement channels through an unsecured communication network. Through a simulation example, we substantiate the effectiveness of our proposed approach. The proposed method emerges as a relevant tool for designing attack sequences tailored for injection into control systems, offering valuable insights for testing detection mechanisms.