Finite frequency distributed fault detection in sensor networks with memory event-triggered scheme and deception...

Area of Science:

• Control Systems Engineering
• Networked Systems Analysis
• Nonlinear System Dynamics

Background:

• Distributed fault detection (FD) is crucial for nonlinear switched systems (SSs) with time-varying delays in sensor networks.
• Network bandwidth limitations and deception attacks pose significant challenges to traditional FD methods.

Purpose of the Study:

• To develop a finite frequency distributed fault detection filter (FDF) for discrete-time nonlinear SSs with time-varying delays.
• To address challenges posed by sensor networks, distributed memory event-triggered schemes (METS), and malicious deception attacks.

Main Methods:

• Utilizing the sojourn probability method to describe the switching mechanism of SSs.
• Employing a distributed memory event-triggered scheme (METS) to reduce network load.
• Designing a finite frequency distributed FD filter (FDF) for each sensor node, considering METS and deception attacks.
• Deriving existence conditions for the FDF using linear matrix inequalities (LMIs).

Main Results:

• The proposed FDF ensures mean-square stability of the augmented filter system via Lyapunov stability analysis.
• The filter meets specified full-frequency H∞ and finite-frequency H- level bounds.
• Comparative simulations confirm superior performance of the finite-frequency distributed FDF over existing full-frequency solutions.

Conclusions:

• The developed finite frequency distributed FDF effectively detects unknown faults in nonlinear SSs within sensor networks.
• The FDF offers enhanced speed and accuracy compared to existing methods, particularly under network constraints and adversarial conditions.