A Dynamic Event-Triggered Approach to Recursive Filtering for Complex Networks With Switching Topologies Subject to...

Area of Science:

• Control Systems Engineering
• Network Science
• Stochastic Systems

Background:

• Complex networks (CNs) exhibit dynamic behaviors like switching topologies and random sensor failures.
• Recursive filtering is crucial for estimating states in such systems but is challenged by network uncertainties.
• Existing methods may not efficiently handle both dynamic topology and sensor issues while optimizing communication.

Purpose of the Study:

• To design dynamic event-triggered recursive filters for nonlinear CNs.
• To address challenges posed by switching topologies and random sensor failures.
• To minimize filtering error covariance while ensuring guaranteed bounds.

Main Methods:

• Utilizing a Markov chain to model topology switching.
• Incorporating a dynamic event-triggered transmission protocol to reduce communication load.
• Employing the induction method to derive and minimize the filtering error covariance upper bound.

Main Results:

• An upper bound on the filtering error covariance was derived.
• The derived upper bound was minimized through optimal filter parameter design.
• The proposed filtering scheme demonstrated effectiveness via a simulation example.

Conclusions:

• The developed dynamic event-triggered filters effectively manage nonlinear CNs with complex dynamics.
• The approach ensures robust filtering performance despite random sensor failures and topology switching.
• This method offers a balance between filtering accuracy and communication efficiency.