Detecting Sabotage Attacks in Additive Manufacturing Using Actuator Power Signatures

Authors:

Abstract
Additive manufacturing (AM), a.k.a. 3D printing is increasingly used to manufacture functional parts of safety-critical systems. The AM's dependence on computerization raises the concern that the AM process can be tampered with, and a part's mechanical properties sabotaged. To address this threat, we propose a novel approach for detecting sabotage attacks based on trusted monitoring of the current delivered to each printer motor. The proposed approach offers numerous advantages: 1) it is non-invasive in a time-critical process, 2) it can be retrofitted in legacy systems, and 3) it can be air-gapped from the computerized components of the AM process, making simultaneous compromise more difficult. We evaluated the approach on five categories of toolpath command-level manipulations that impact the geometry of the 3D printed object. Our evaluation showed that all but one tested category of attacks can be reliably detected, even if a single toolpath command is modified.
Funding Information
  • National Institute of Standards and Technology (70NANB18H266)
  • National Institute of Standards and Technology (70NANB18H266)
  • National Science Foundation (1624944)

This publication has 20 references indexed in Scilit: