Detectability Analysis of Cyber Attacks on CPSs

Security of Cyber-Physical Systems (CPS) against cyber attacks is an important yet challenging problem. Since most cyber attacks happen in erratic ways, it is difficult to describe them systematically. In our research, instead of identifying a specific cyber attack model, we are focused on analyzing the system’s response during cyber attacks. Considering the case in which intelligent cyber attackers avoid detection by the monitoring system with carefully designed attacks, our main objective is to investigate the performance of such stealthy cyber attacks from the system’s perspective. We investigate three kinds of stealthy cyber attacks according to the attacker’s ability to compromise the system. Based on the information about the dynamics of the system and existing hypothesis testing algorithms, we derive the necessary and sufficient conditions under which the attacker could perform cyber attacks without being detected. The threat of these cyber attacks is illustrated via an Unmanned Aerial Vehicle (UAV) navigation example. Figure 3 demonstrates the effects of three different stealthy cyber attacks on the 3-D tracking of the UAV. Figure 4 compares the values of the residual statistics of the nominal system and the system under the attacks shown in Figure 3. This demonstrates that the residuals under attack have no significant change in their statistical properties, which implies the attacks are undetectable via the residual tests. Thus, the estimated UAV trajectories under the attacks can significantly deviate from the actual UAV trajectory without being detected, i.e., the UAV can be hijacked to any place that the attacker wants.

Figure 3. Trajectory estimates under three different cyber attacks

Figure 4. Test statistics under three different cyber attacks