Resilient State and Input Estimation for Complex Network Subject to Cyber Attack: A Set-Membership Method.

Area of Science:

• Control Systems Engineering
• Network Security
• Signal Processing

Background:

• Complex networks are vulnerable to malicious attacks on sensor communication channels, compromising signal integrity.
• Existing state estimation methods struggle with corrupted input and output signals, necessitating robust solutions.
• Direct measurement of unknown inputs is often impractical in real-world applications.

Purpose of the Study:

• To develop a resilient state estimation framework for complex networks under sensor attacks.
• To jointly estimate states and unknown inputs, mitigating the impact of corrupted signals.
• To ensure bounded state estimation errors even with malicious input and output signal interference.

Main Methods:

• A novel set-membership filter framework integrated with an unknown input estimator was designed.
• Rigorous mathematical induction was used to derive conditions for bounded state estimation.
• Filter gains were determined to decouple state estimation error from unknown inputs and minimize estimation bounds.

Main Results:

• The proposed framework guarantees that state estimation errors remain within a defined ellipsoid under malicious attacks.
• The method effectively mitigates the adverse effects of corrupted input and output signals.
• The framework was successfully extended to networks with direct feedthrough, demonstrating its versatility.

Conclusions:

• The developed set-membership filter provides a resilient solution for state estimation in complex networks susceptible to sensor attacks.
• The joint estimation of states and unknown inputs enhances robustness against signal corruption.
• Numerical simulations and experimental validation confirm the effectiveness and resilience of the proposed estimators.