Unreliable networks with random parameter matrices and time-correlated noises: distributed estimation under deception...

Area of Science:

• Control Systems and Signal Processing
• Networked Systems Analysis
• Estimation Theory

Background:

• Networked systems are susceptible to various uncertainties, including random parameters, correlated noises, and deception attacks.
• Accurate state estimation is crucial for the reliable operation of these systems.
• Existing methods often struggle to address multiple network-induced phenomena simultaneously.

Purpose of the Study:

• To develop robust distributed filtering and fixed-point smoothing algorithms for networked systems.
• To provide a unified framework for estimation under incomplete information and various uncertainties.
• To address challenges posed by random parameter matrices, time-correlated additive noises, and random deception attacks.

Main Methods:

• A two-stage distributed estimation algorithm is proposed.
• The first stage generates intermediate estimators using local and adjacent node data.
• The second stage combines intermediate estimators via least-squares matrix-weighted linear combinations, utilizing a covariance-based technique without requiring signal evolution models.

Main Results:

• The developed algorithms effectively handle random parameter matrices, time-correlated additive noises, and random deception attacks.
• A unified framework is established for distributed estimation in systems with incomplete information.
• Numerical experiments validate the algorithms' applicability and effectiveness, demonstrating the impact of uncertainties and attacks on estimation accuracy.

Conclusions:

• The proposed distributed estimation algorithms offer a robust solution for networked systems under complex uncertainties.
• The covariance-based approach provides flexibility by not requiring signal evolution models.
• The study highlights the critical need for robust estimation strategies in the presence of network vulnerabilities.