Title:
EFFICIENT LEARNING FOR HARDWARE SECURITY VALIDATION USING ELECTROMAGNETIC SIDE CHANNELS
EFFICIENT LEARNING FOR HARDWARE SECURITY VALIDATION USING ELECTROMAGNETIC SIDE CHANNELS
Author(s)
Jorgensen, Erik J.
Advisor(s)
Zajić, Alenka
Bloch, Matthieu R.
Bloch, Matthieu R.
Editor(s)
Collections
Supplementary to
Permanent Link
Abstract
The objective of this thesis is to combine the non-destructive monitoring advantages of standard and backscattering electromagnetic side channels with modern machine learning techniques to efficiently validate the authenticity of individual integrated circuits installed on a motherboard.
The authenticity of integrated circuits is of increasing concern as more steps in the device manufacturing supply chain are outsourced, especially in light of severe global semiconductor shortages. Common methods for integrated circuit validation rely on either destructive techniques before high resolution imaging of the circuit interconnects or functional testing of a variety of test inputs with automated test equipment. These methods are time-consuming or even intractable to detect counterfeit components or stealthy modifications of their underlying circuitry.
Side channels are any means of remotely leaking information related to a circuit's activity or architecture. Our work takes advantage of the electromagnetic (EM) side channel to remotely capture identifying information emitted from or backscattered off integrated circuits in the form of EM signals that can be used to validate their authenticity.
This research attempts to alleviate the need for time-consuming and expensive destructive validation methods for hardware security by robustly detecting inauthentic or modified integrated circuits with remote EM side-channel measurements. The first aim of this research is to apply deep learning methods to classify and detect counterfeits of major ICs on a variety of motherboards. The second aim is to leverage hyperspectral scanning with the backscattered EM side-channel and a novel active learning method to detect dormant hardware trojans several times smaller than before. The last aim is to develop a compressed sensing approach to heavily reduce sampling for hardware trojan detection as well as to develop a hyperspectral characterization of expected and anomalous circuits.
Sponsor
Date Issued
2022-08-15
Extent
Resource Type
Text
Resource Subtype
Dissertation