Active Bayesian Inference for Robust Control under Sensor False Data Injection Attacks

2026-04-13

AI Summary

This work addresses the challenge of inaccurate state estimation in cyber-physical systems caused by false data injection attacks. The authors propose a unified framework that integrates active probing with Bayesian inference to dynamically identify compromised sensors and maintain robust control. By modeling the sensing process as a bipartite graph and constructing a Bayesian network incorporating anomaly alerts, the method enables real-time inference of sensor integrity. Leveraging system nonlinearities, an active probing strategy is designed to enhance the distinguishability among competing attack hypotheses, allowing selective disabling of corrupted nodes. Notably, this approach achieves progressively improving attack identifiability and reliable state estimation even during ongoing attacks. Experimental results on an inverted pendulum system under both single- and multi-sensor attacks demonstrate significant performance gains over baseline methods based on outlier robustness and prediction, with particularly pronounced advantages in long-duration attack scenarios.

Technology Category

• Intelligent Robots: State Estimation
• Reasoning under Uncertainty: Probabilistic Inference
• Cognitive Modeling & Cognitive Systems: Conceptual Inference and Reasoning

Application Category

• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Security and Privacy: Large-scale security measurements

Abstract

We present a framework for bridging the gap between sensor attack detection and recovery in cyber-physical systems. The proposed framework models modern-day, complex perception pipelines as bipartite graphs, which combined with anomaly detector alerts defines a Bayesian network for inferring compromised sensors. An active probing strategy exploits system nonlinearities to maximize distinguishability between attack hypotheses, while compromised sensors are selectively disabled to maintain reliable state estimation. We propose a threshold-based probing strategy and show its effectiveness via a simplified partially observable Markov decision process (POMDP) formulation. Experiments on an inverted pendulum under single and multi-sensor attacks show that our method significantly outperforms outlier-robust and prediction-based baselines, especially under prolonged attacks.

Problem

• sensor false data injection attacks
• cyber-physical systems
• robust control
• state estimation
• attack detection

Innovation

• Active Bayesian Inference
• Sensor Attack Detection
• Bipartite Perception Graph
• Anomaly-aware Probing
• Robust State Estimation