Date(s) - 10 Nov 2021
1:10 PM - 2:00 PM
Speaker: Varghese Mathew Vaidyan, ECpE Graduate Student
Advisor: Akhilesh Tyagi
Title: Control System Processors and Electromagnetic Failure Analysis
Abstract: Control system processors in networked control systems malfunction due to adverse operating conditions and are prone to malware attacks. Distance monitoring of stuck-at flaws and malware is more reliable and easier to integrate into IoT platforms. A novel failure/performance analysis and Machine Learning (ML) modeling using an Electromagnetic (EM) Spectral domain-based analysis framework that is fully independent of the controller processor is presented here. It can evaluate the causes of failure/performance issues even on a sophisticated 6-stage pipelined microarchitecture without modifications of the monitored system. Because standard malware cannot traverse EM side-channels, our monitor is impervious to malware attacks that have already infected the controller. It can also work without requiring any changes to the operational controller. The electromagnetic traces from the controller in the frequency domain are examined using Machine Learning (ML) models of stuck-at failures and infections. This machine learning model is tested using Support Vector Machines (SVM), and other ML classifiers. Our research on controller implementations employing a 6-stage pipelined ARM Cortex-M7 has shown that a control system stuck-at problems and malware can be distinguished with very high accuracy.
Bio: Varghese Mathew Vaidyan received the M.S. degree in electronics and electrical engineering from University of Glasgow. He is currently pursuing the Ph.D. degree with the Department of Electrical and Computer Engineering, Iowa State University. His research interests include Electromagnetic fault monitoring methods and Machine Learning.