Holistic Control and Management of Industrial Wireless Processes

We address the problem of cyber security of process control systems. We consider the context of industrial control systems in which an attacked sensor node sends false measurements of the system state, resulting in damage to the physical plant. We employ a "dynamic watermarking" technique to secure process control systems against malicious attacks. First, our simulation results demonstrate that this method successfully detects a malicious activity in time to prevent overflows in coupled water tanks. Second, in order to exhibit its applicability to more complex and practical industrial systems, we show by simulation study that dynamic watermarking can protect the Tennessee Eastman Process system by detecting a malicious sensor attack.