CPS: Breakthrough: Towards a Science of Attack Composition, Mitigation, and Verification in Cyber-Physical Systems: A Passivity-Based Framework
Lead PI:
Radha Poovendran
Co-Pi:
Abstract
This project focuses on modeling and mitigating cyber attacks on Cyber-Physical Systems (CPS), which are increasingly prevalent in all aspects of society such as health care, energy, and transportation. Attacks initiated on the cyber components of CPS can be mounted remotely at little economic cost and can significantly degrade the safety and performance of CPS due to the tight coupling between cyber and physical components. This project develops a passivity-based framework for modeling, composing, and mitigating multiple attacks on CPS. Passivity is an energy dissipation property that provides basic rules for analyzing and composing interconnected systems. In addition to passive adversary models and composition rules, this project will investigate techniques for decomposition of composed attack models into basic primitives which will lead to development of new mitigation strategies. Approximate bi-simulation techniques will be introduced to verify the developed adversary models and mitigation strategies. The proposed approach is general and will be applicable to mitigate CPS security challenges arising in multiple sectors including transportation, energy, manufacturing, and others. The goals of the project are as follows: (a) research and development of passive dynamical models of multiple attacks, as well as characterization of the class of attacks that admit a passive representation; (b) investigation and development of passivity-based composition and decomposition rules, enabling identification of new attack variants and associated mitigation strategies; (c) research and development of approximate techniques for verification of composed adversary models and mitigation strategies; and (d) validation and prototyping of the proposed models through an experimental testbed.
Radha Poovendran
Performance Period: 10/01/2014 - 09/30/2017
Institution: University of Washington
Sponsor: National Science Foundation
Award Number: 1446866