Foundations of a CPS Resilience - July 2021

PI: Xenofon Koutsoukos

HARD PROBLEM(S) ADDRESSED

The goals of this project are to develop the principles and methods for designing and analyzing resilient CPS architectures that deliver required service in the face of compromised components. A fundamental challenge is to understand the basic tenets of CPS resilience and how they can be used in developing resilient architectures. The primary hard problem addressed is resilient architectures. In addition, the work addresses scalability and composability as well as metrics and evaluation. 

PUBLICATIONS

[1]    Feiyang Cai, Ali Ozdagli, Xenofon Koutsoukos, “Detection of Dataset Shifts in Learning-Enabled Cyber-Physical Systems using Variational Autoencoder for Regression”, IEEE International Conference on Industrial Cyber-Physical Systems (ICPS 2021). May 10-12, 2021.

[2]    Ajay Chhokra, Carlos Barreto, Abhishek Dubey, Gabor Karsai and Xenofon Koutsoukos, “Power-Attack: A comprehensive tool-chain for modeling and simulating attacks in power systems”, 9th Workshop on Modeling and Simulation of Cyber-Physical Energy Systems (MSCPES 2021). May 18, 2021.

[3]    Dimitrios Boursinos and Xenofon Koutsoukos, “Assurance Monitoring of Learning Enabled Cyber-Physical Systems using Inductive Conformal Prediction based on Distance Learning”, Artificial Intelligence for Engineering Design, Analysis and Manufacturing, 35(2), 251-264, May 31, 2021.

[4]    Waseem Abbas, Mudassir Shabbir, Yasin Yazıcıoğlu and Xenofon Koutsoukos, "Edge Augmentation With Controllability Constraints in Directed Laplacian Networks," IEEE Control Systems Letters. Published online: June 14, 2021. 
 

KEY HIGHLIGHTS

This quarterly report presents two key highlight that demonstrate: (1) an algorithm for edge augmentation in networked multi-agent systems for improving controllability, a key property for robustness and resilience and (2) an approach for assurance monitoring of learning-enabled CPS.

Highlight 1: Edge augmentation in networked multi-agent systems

In a networked multi-agent system, a frequent approach to improve network connectivity is to systematically increase interconnections between agents. Edge augmentation is useful for improving network connectivity, robustness, and resilience, but on the other hand, adding edges could adversely impact network controllability. We study the problem of maximum edge augmentation in a directed network of agents with Laplacian dynamics while preserving the controllability specification. We consider the network’s strong structural controllability (SSC), which depends (apart from the set of input nodes) only on the structure of the underlying graph defined by the edge set of the graph. We utilize two widely used bounds that are based on the ideas of Zero Forcing (ZF) and distances between nodes in graphs. We present an optimal edge augmentation algorithm for adding the maximum number of edges in a directed graph while preserving the ZF-based bound on the dimension of SSCS. We also discuss edge augmentation in graphs that preserves the distance-based bound on the dimension of SSCS. Finally, we provide a randomized algorithm that adds maximal edges in a directed graph while preserving the distance-based bound on the dimension of SSCS. Our results are presented in [1]. 

[1]    Waseem Abbas, Mudassir Shabbir, Yasin Yazıcıoğlu and Xenofon Koutsoukos, "Edge Augmentation with Controllability Constraints in Directed Laplacian Networks," IEEE Control Systems Letters. Published online: June 14, 2021.

Highlight 2: Assurance Monitoring of Learning-enabled CPS

Machine learning components such as deep neural networks are used extensively in cyber-physical systems (CPS). However, such components may introduce new types of hazards that can have disastrous consequences and need to be addressed for engineering trustworthy systems. Although deep neural networks offer advanced capabilities, they must be complemented by engineering methods and practices that allow effective integration in CPS. We proposed an approach for assurance monitoring of learning-enabled CPS based on the conformal prediction framework. In order to allow real-time assurance monitoring, the approach employs distance learning to transform high-dimensional inputs into lower size embedding representations. By leveraging conformal prediction, the approach provides well-calibrated confidence and ensures a bounded small error rate while limiting the number of inputs for which an accurate prediction cannot be made. We demonstrate the approach using three datasets of mobile robot following a wall, speaker recognition, and traffic sign recognition. The experimental results demonstrate that the error rates are well-calibrated while the number of alarms is very small. The method is computationally efficient and allows real-time monitoring. Our results are presented in [2]. Current and future work includes using the approach for detection and classification of attacks in CPS/IoT. 

[2]    Dimitrios Boursinos and Xenofon Koutsoukos, “Assurance Monitoring of Learning Enabled Cyber-Physical Systems using Inductive Conformal Prediction based on Distance Learning”, Artificial Intelligence for Engineering Design, Analysis and Manufacturing, 35(2), 251-264, May 31, 2021.

COMMUNITY ENGAGEMENTS

Meeting with Chris Hankin (and Perry Alexander) and discuss the synergies between the lablets and Research Institute in Trustworthy Inter-connected Cyber-Physical Systems (RITICS).