ROSIS: Resilience Oriented Security Inspection System against False Data Injection Attacks

Current radar-based Air Traffic Service (ATS) providers lack the preservation of privacy for airspace operations of selected flight plans, positions, and state data; requiring security assurance. Recent events have shown that modern unmanned aerial vehicles (UAVs) are vulnerable to attack and subversion through software flaws or sometimes malicious devices that are present on urban air mobility (UAM) communication networks, which increases the need for cyber awareness including the risk of cyber intrusion. Many stealthy attacks (such as False Data Injection Attacks (FDIAs)) are hard to detect on avionics systems, as they can compromise measurements from sensors and bypass the sensor's basic “faulty data” detection mechanism and remain undetected. Such attacks on a UAM system may not even present their impact but propagate from the sensor to fool the system by predicting a delayed asset failure or maintenance interval. In this paper, we develop a Resilience Oriented Security Inspection System (ROSIS) to maximize UAM capability to secure data accessing and sharing among aircraft and Air Traffic Service (ATS) service providers. Specifically, we collect and demonstrate the effect of generalized FDIAs on wireless sensors of a turbofan engine using NASA's C-MAPSS simulator, and develop data-driven based deep learning methods (Long Short-Term Memory (LSTM) and Gated Recurrent Unit (GRU)) for detecting abnormal features. A graphical physics-informed Bayesian Network model is developed to represent the dynamic nature of the engine to predict health conditions accordingly. The ROSIS model characterizes the condition-symptom relationships of different engine components and sensors. A hybrid software-in-the-loop (SITL) and hardware-in-the-loop (HITL) design is also developed to evaluate the effectiveness of the ROSIS defense mechanisms. Our experiments validate the performance of ROSIS in detection accuracy and efficiency against FDIAs.