TUM Logo

Exploring channels for anomaly detection in additive manufacturing

Exploring channels for anomaly detection in additive manufacturing

Supervisor(s): Nikolai Puch
Status: finished
Topic: Anomaly Detection
Author: Petar Zoric
Submission: 2022-10-17
Type of Thesis: Masterthesis
Thesis topic in co-operation with the Fraunhofer Institute for Applied and Integrated Security AISEC, Garching


Additive manufacturing has gained a lot of popularity and is used in a variety of fields. Even safety-critical

applications like medical devices are implementing 3D-printed parts. However, cyber-physical systems like

3D printing environments can get targeted by attackers. When it comes to cyber-physical systems, an important

but often overlooked property are side channels, like sound or heat emissions. This work looks at several side

channels that are given in additive manufacturing systems.

The thesis provides an in-depth analysis of four of the most important channels and several side channel combinations

from the perspective of an attacker and a defender. The concept of maximum restorable information is introduced and

is used throughout the thesis to determine the effects of certain attacks and whether they can be detected or reconstructed

by a side channel.

After performing the analysis, the acoustic and the spatial side channel are analyzed further practically. A prototype for

gathering and analyzing acoustic data is developed and the feasibility of nozzle tracking with home-grade equipment is

tested. Based on the result of the acoustic analysis, the assumption that specific nozzle movements lead to unique acoustic

profiles is further manifested.