Event-based distributed filtering against deception attacks for sensor networks with quantization effect.

Area of Science:

• Control Systems Engineering
• Networked Systems Security
• Signal Processing

Background:

• Distributed systems are vulnerable to uncertainties, model mismatch, and sophisticated deception attacks.
• Bandwidth limitations and random attacks pose significant challenges to secure filtering in networked systems.

Purpose of the Study:

• To design time-varying filters for distributed systems under uncertainty and two-stage random deception attacks.
• To guarantee H-infinity secure performance over a finite time horizon despite communication constraints.

Main Methods:

• Utilized a quantization strategy to reduce communication load.
• Implemented an event-triggered mechanism for efficient data transmission.
• Derived sufficient conditions using recursive matrix inequalities for filter existence.

Main Results:

• Developed time-varying filters ensuring guaranteed H-infinity secure performance.
• Successfully addressed challenges of uncertainty, model mismatch, and random deception attacks.
• Demonstrated filter validity through simulations on a numerical example and a continuous-stirred tank reactor system.

Conclusions:

• The proposed methodology effectively guarantees secure filtering performance in complex, uncertain, and attacked distributed systems.
• Event-triggered and quantization strategies are crucial for efficient and secure networked control.
• The derived conditions and online filter gain computation offer a robust solution for practical applications.