Online Monitoring of a Cyber Physical System Against Control Aware Cyber Attacks

In various industrial plants like power and chemical plants, the system operation is controlled by embedded controller(s). Any intentional malfunction of a critical controller can lead to shut down or failure of vital parts. It further leads the control plant into unsafe mode. Malware attacks can result in tremendous cost to the organization in terms of cleanup activity. Process related threats occur when the attacker gains control of the system and performs unintended actions. Industrial plants being a complex system need a wholesome approach for attack detection and prevention. In this paper we propose to use a geometric method to detect anomaly in a control system behavior which can possibly indicate a malware attack.The paper focuses on analysis of the large data sets for anomaly detection by using computational geometric methods to observe and analyze trends in the controller's output.