Systematic evaluation of abnormal detection methods on gas well sensor data

IADF-CPS: Intelligent Anomaly Detection Framework towards Cyber Physical Systems

“…Furthermore, no prior work in detecting coordinated cyber attacks in CRAS cyber se- Inference Attacks Communication links and Sensor Nodes UNSW NB15 and power system dataset Smart Power Networks 92% of accuracy is obtained [22] Probe attack, DoS attack, and unauthorized access attacks Communication links and controllers NSL-KDD dataset Industrial Control Networks 97.8% of accuracy is obtained [26] Phasor Measurement Communication links and Controllers Simulated IEEE 9 bus Smart grids 94.1% is obtained [27] Unspecified cyber attacks Sensor nodes and Actuator nodes Simulated data from gas turbines Smart grids FPR rate of 0.000006 is obtained [28] Fuzzy attack, data spoofing, and exploits attacks Actuator nodes, Communication links, and Sensor nodes Car hacking dataset and UNSW-NB15 Internet of Vehicles 99% of accuracy is obtained [29] Replay attacks Controllers 118 bus systems Smart grids MAPE obtained as 3.51% [30] False data injection attacks Actuator nodes and Sensor nodes SWaT dataset Water Treatment Plant 89% of accuracy is obtained curity has been done. Authors in [34] have proposed IDS based on statistical and rule mining methods. Authors defined some set of rules for particular features of network such as flow layer, inter-flow layer, and packet layer.…”