Collaborative Research: Cyber-secure and Resilient Supervisory Control of Networked Discrete-Event Systems

Wayne State University, Detroit MI

Investigators

Abstract

With the rapid development of computer, communication, and control technologies, modern engineering systems consist of many distributed and networked physical plants, control units, and other devices. These networked systems are vulnerable to cyber attacks. The principal objective of this project is to develop formal methodologies to synthesize control logic for cyber-secure and resilient control of networked dynamic systems subject to attacks on their communication, sensing, or actuation infrastructure. By providing formal methodologies to detect cyber attacks and synthesize provably-resilient control logic for networked systems, this work will enhance the reliability and performance of advanced control systems that embed complex control logic in their control structure. The developments in this project will be model-based, thereby ensuring their applicability to a large class of technological networked systems subject to cyber attacks. In particular, power grids with high penetration of renewable energy sources and electric vehicles will be considered, a technological area of considerable societal importance. The scientific foundation of the proposed approach to synthesizing cyber-secure and resilient control logic at the supervisory layer of the control system is the theory of control and diagnosis of event-driven systems developed in control engineering. Networked control systems will be modeled in the framework of transition systems with discrete state spaces and event-driven dynamics, or discrete event systems. Supervisory control at the higher layer of complex control systems will be designed to ensure that the discrete model satisfies a set of given logical specifications related to safety properties and liveness properties and are resilient to cyber attacks on the communication infrastructure or on sensors and actuators themselves. To detect cyber attacks and other faults, the theory of diagnosability of discrete event systems will be extended to networked systems subject to joint sensor and actuator attacks. The research plan blends supervisory control techniques from formal methods in control engineering with algorithmic techniques from discrete games in theoretical computer science. In addition to the theoretical and algorithmic components of this project, the researchers will investigate the application area of smart power grids with high penetration of renewables and electric vehicles, where the control and diagnosis problems formulated herein are highly relevant. Work in this application area will involve modeling, testing of the synthesis methodologies developed on these models, and construction of a simulation testbed for performance evaluation of the developed methodologies. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.