From Passive Monitoring to Active Defense: Resilient Control of Manipulators under Cyberattacks
Gabriele Gualandi, Alessandro Vittorio Papadopoulos
AI summary
Problem
Stealthy false data injection attacks bypass passive anomaly detectors and exploit the integrator vulnerability in feedback-linearized manipulators, causing large end-effector deviations without triggering alarms.
Approach
The method attenuates control inputs using a monotonic function of an anomaly score generated by a novel actuation-projected, measurement-free state predictor, ensuring closed-loop stability and probabilistic performance guarantees.
Key results
- Formalization of stealthy FDIAs revealing an integrator vulnerability and reducing optimal attack synthesis to a convex QCQP
- Development of anomaly-aware command scaling based on a measurement-free actuation-projected state predictor
- Proof of probabilistic guarantees on bounded actuation loss and closed-loop stability under the defense
- Simulation validation on a 6-DOF manipulator showing significant reduction in attack-induced deviations while preserving nominal performance
Why it matters
Provides a practical, mathematically grounded resilience mechanism for networked robotic systems vulnerable to stealthy cyberattacks.
Abstract
Cyber-physical robotic systems are vulnerable to false data injection attacks (FDIAs), in which an adversary corrupts sensor signals while evading residual-based passive anomaly detectors such as the χ2 test. Such stealthy attacks can induce substantial end-effector deviations without triggering alarms. This paper studies the resilience of redundant manip- ulators to stealthy FDIAs and advances the architecture from passive monitoring to active defence. We formulate a closed- loop model comprising a feedback-linearized manipulator, a steady-state Kalman filter, and a χ2-based anomaly detector. Building on this passive monitoring layer, we propose an active control-level defence that attenuates the control input through a monotone function of an anomaly score generated by a novel actuation-projected, measurement-free state predictor. The pro- posed design provides probabilistic guarantees on nominal actuation loss and preserves closed-loop stability. From the attacker perspective, we derive a convex QCQP for computing one-step optimal stealthy attacks. Simulations on a 6-DOF pla- nar manipulator show that the proposed defence significantly reduces attack-induced end-effector deviation while preserving nominal task performance in the absence of attacks.