Distributed Adaptive Probabilistic Event‐Triggered Filtering for Fuzzy Systems Under Data‐Important‐Aware DoS Attack

Abstract

This study proposes a novel distributed filtering framework for discrete‐time nonlinear systems over sensor networks, tackling challenges from bandwidth limitations, communication delays, and cyber‐attacks. An adaptive probabilistic event‐triggered mechanism is introduced, utilizing a dynamic threshold with probabilistic delay division to reduce transmissions and handle stochastic delays. This is combined with a dynamic saturation function featuring time‐varying adaptively bounds to suppress outliers and cyber‐attacks. A resilient distributed Takagi–Sugeno (T–S) fuzzy filtering strategy is then developed based on piecewise Lyapunov functionals and linear matrix inequalities, ensuring stochastic stability and prescribed performance. Additionally, a data‐importance‐aware denial‐of‐service (DoS) attack model is proposed, where attackers target high‐value packets via a weighted error metric to maximize impact efficiently. Simulations confirm that the framework improved estimation accuracy, lowers communication rates, and strengthens resilience against targeted attacks compared to existing methods. The integrated design offers a theoretically guaranteed and practically effective solution for secure distributed filtering in cyber‐physical systems.