Dynamic Output Feedback Linear Quadratic Control for CPSs Under Sparse Attacks.

Area of Science:

• Control Systems Engineering
• Cyber-Physical Systems Security
• Robust Control Theory

Background:

• Cyber-physical systems (CPSs) are increasingly vulnerable to sophisticated actuator and sensor attacks.
• Ensuring system stability and performance under adversarial conditions is a critical challenge.
• Existing control methods may struggle with the computational demands of real-time state estimation during attacks.

Purpose of the Study:

• To propose a novel dynamic output feedback (DOF) linear quadratic (LQ) control strategy for CPSs under sparse attacks.
• To develop efficient online attack detection mechanisms to ensure system stability.
• To reduce the computational complexity compared to traditional data-based optimal control methods.

Main Methods:

• Transformation of the CPS into hybrid systems based on potential attack scenarios.
• Analysis of the DOF LQ (dLQ) control scheme, including similarity transformations and observable saddle points.
• Development of two online attack detection mechanisms: Adaptive Switching Mechanism (ASM) and Improved ASM (IASM).

Main Results:

• The proposed hybrid control scheme guarantees asymptotic stability of the closed-loop system.
• IASM demonstrates faster attack detection capabilities compared to ASM.
• The method significantly reduces computational complexity by requiring state solutions only during control mode switching.

Conclusions:

• The presented DOF LQ control strategy with integrated ASM/IASM offers a robust and computationally efficient solution for securing CPSs against sparse attacks.
• The approach effectively balances control performance with enhanced security.
• Simulation examples validate the effectiveness and superiority of the proposed method.