Fundamental Connections between Opacity and Attack Detection in Linear Systems

Varkey M. John




Research Conducted

Title: On Connections between Opacity and Security in Linear Systems Our research focussed on finding a relation between attack detection and the privacy notion ‘opacity’ in linear systems. We showed that a fundamental trade-off exists between these two properties. In particular, if an opaque system is subjected to attacks, all attacks cannot be detected. We first characterized the opacity conditions for the system in terms of its weakly unobservable subspace (WUS) and showed that the number of opaque states is proportional to the size of the WUS. Further, we established conditions under which increasing the opaque sets also increases the set of undetectable attacks. This highlighted a fundamental trade-off between security and privacy. We demonstrated application of our results on a remotely controlled automotive system. This work was submitted to the IEEE Conference on Decision and Control (CDC), 2022. The preprint is available here: https://arxiv.org/abs/2206.06074.

Other Work

The CNI website is currently hosted using WordPress. However, this has issues such as reliability, security and maintainability. Therefore, it was proposed to develop the CNI website using a static site generator instead so as to meet these requirements. For this purpose, we developed the entire CNI website from scratch in Jekyll framework. Additional custom features were also incorporated into the website. The design and layout of the entire website were also improved tremendously. The new website was hosted on GitHub: https://cni-iisc.github.io/cni-website/index.html