Resilient estimation and safe control for cyber-physical systems

Wan, Wenbin

Permalink

https://hdl.handle.net/2142/117649 Copy

Description

Title

Resilient estimation and safe control for cyber-physical systems

Author(s)

Wan, Wenbin

Issue Date

2022-11-21

Director of Research (if dissertation) or Advisor (if thesis)

Hovakimyan, Naira

Doctoral Committee Chair(s)

Hovakimyan, Naira

Committee Member(s)

• Sha, Lui
• Salapaka, Srinivasa
• Voulgaris, Petros
• Kim, Hunmin

Department of Study

Mechanical Sci & Engineering

Discipline

Mechanical Engineering

Degree Granting Institution

University of Illinois at Urbana-Champaign

Degree Name

Ph.D.

Degree Level

Dissertation

Keyword(s)

• Cyber-physical Systems
• Resilient Estimation
• Robust Control
• Adaptive Control
• Interval Estimation

Language

eng

Abstract

The recent decade has been critical in designing and deploying cyber-physical systems (CPS). CPS Security and CPS safety often are essential. The research proposed in this dissertation aims to enable safe operation for cyber-physical systems (CPS) subject to significant uncertainties, such as malicious attacks, unforeseen environments, and model uncertainties, by integrating resilient estimation algorithms and safe control methods. First, we consider the problem of a safety-constrained control architecture design against GPS spoofing/jamming attacks. We develop a resilient estimation algorithm to detect attacks and design control algorithms based on the model predictive controller (MPC) subject to limited sensor availability due to the sensor attacks. In another scenario of actuator attacks, we propose a constrained attack-resilient estimation algorithm (CARE) against the CPS attacks. The CARE can simultaneously estimate the compromised system states and the attack signals. In particular, CARE first provides minimum-variance unbiased estimates and then projects the estimates onto the constrained space induced by physical constraints and operational limitations. The proposed CARE performs better in estimation and attack detection by reducing estimation errors, covariances, and false negative rates. Following that, we extend our resilient estimation algorithm to a spatio-temporal framework. Building on the proposed resilient spatio-temporal filtering, we design a proactive adaptation architecture for connected vehicles in unforeseen environments, synthesizing techniques in spatio-temporal data fusion and robust adaptive control. Finally, we propose an efficient interval estimation method for estimating systems under faulty model uncertainties. The method applies to a broad class of systems with a large uncertainty setup.

Graduation Semester

2022-12

Type of Resource

Thesis

Handle URL

https://hdl.handle.net/2142/117649

Copyright and License Information

Copyright 2022 Wenbin Wan

Owning Collections