Adaptive Detection Method for Physical Access Attack against Temperature Interference

Industrial Control System (ICS) has been widely used in the Cyber-Physical System (CPS). Traditional ICS mainly adopts network isolation to prevent network intrusion. However, adversaries can attack the ICS via physically attaching devices into its underlying field bus networks. Moreover, this physical access attack is difficult to detect by traditional traffic-based intrusion detection methods. We have proposed a physical access detection method to detect the attack. However, the current intrusion detection method has ignored the impact of environment factors such as the ambient temperature, whose change would cause high cost and poor detection performance. To address this issue, we propose an adaptive detection method for physical access attack against temperature interference. First, we establish a temperature-sensitive fingerprint model and analyze the physical access attack under temperature interference. Then, we extract features for detection using the time-domain characteristics of the fingerprint signal. After that, we generate an adaptive threshold model by fitting the detection threshold with the ambient temperature. Simulation results demonstrate that our proposed adaptive detection method could effectively detect physical access attacks under different temperatures.