Data-driven Cyberattack Synthesis against Network Control Systems

TL;DR

This paper presents a novel data-driven method for synthesizing hybrid cyberattacks on Network Control Systems by leveraging system modeling and reachable sets, demonstrated through multi-aerial vehicle control scenarios.

Contribution

It introduces a white-box attack synthesis technique combining eavesdropping, system modeling, and reachable set analysis for realistic cyberattack generation against NCSs.

Findings

Effective attack synthesis using system models and reachable sets.

Real-time attack directives for NCS agents.

Demonstrated on multi-aerial vehicle formation control.

Abstract

Network Control Systems (NCSs) pose unique vulnerabilities to cyberattacks due to a heavy reliance on communication channels. These channels can be susceptible to eavesdropping, false data injection (FDI), and denial of service (DoS). As a result, smarter cyberattacks can employ a combination of techniques to cause degradation of the considered NCS performance. We consider a white-box cyberattack synthesis technique in which the attacker initially eavesdrops to gather system data, and constructs equivalent system model. We utilize the equivalent model to synthesize hybrid cyberattacks -- a combination of FDI and DoS attacks against the NCS. Reachable sets for the equivalent NCS model provide rapid, real-time directives towards selecting NCS agents to be attacked. The devised method provides a significantly more realistic approach toward cyberattack synthesis against NCSs with unknown parameters. We demonstrate the proposed method using a multi-aerial vehicle formation control scenario.