A resilient framework for sensor-based attacks on cyber–physical systems using trust-based consensus and self-triggered control

Abstract

In networked Cyber–Physical Systems, fidelity of sensor data is critical for the safe operation of the physical plant. Misrepresentation of the data fed back by a primary sensor may cause service disruption or damage to the wider plant through interconnected physical processes. A framework for the recovery of control after an attack has been detected is developed with the following key concepts: (I) a collection of secure secondary sensors that measure energy emissions from components of the system to provide an assessment of the primary sensors’ trust status; (II) a trust-weighted consensus algorithm that fuses estimates from observers of interconnected processes with information from the secondary sensors to accurately reconstruct the state of the attacked process; and (III) a communication-aware self-triggered control protocol that regulates the attacked process using the consented estimate in the absence of reliable data from its primary sensor. The framework is physically tested on a two-dimensional heat conduction system with two kinds of secondary sensors.