Professor awarded $425,000 grant to help prevent sabotage within additive manufacturing
Published: Oct 21, 2021 8:00 AM
By Joe McAdory
Sabotage attacks against 3D printed, additively manufactured components can possibly lead to disastrous mishaps with systems using these components, including but not limited to airplanes. Consequently, it could cause injuries, disrupt industrial production, or slow supply chains.
Mark Yampolskiy, assistant professor in computer science and software engineering at the Samuel Ginn College of Engineering, is working to develop systems that will:
* Detect such attacks
* Localize introduced deviations within the 3D printed parts, and
* Reconstruct deviations for further investigation
His project, “Side-Channel-based Detection, Localization, and Investigation of Sabotage Attacks in Manufacturing,” was recently awarded a $425,000 grant from the National Institute of Standards and Technology (NIST). The proposal builds on prior work that demonstrated the effectiveness of the power side-channel in reconstructing additive manufacturing systems’ behavior and expanding this into a novel forensic tool used to isolate and visualize sabotaged areas.
Side-channels refer to unintentional emanations of a system, or hardware, rather than the program directly. “Side-channels can be measured in the physical domain. Examples are power, electromagnetic, acoustic, and so on,” Yampolskiy explained.
“These channels give hints about what is going on in the cyber-physical world. 3D printers are Cyber-Physical Systems (CPS), meaning that they operate in the physical domain while all their actions are controlled from cyber domain. For a specific digital design file, programs in the cyber domain decide as to which actions must be physically executed by the machine’s components. The final transitions between cyber and physical domains are done by motor controllers that actually transform digital input into physical output of electrical signals, and in our case, the actuators.”
Changes in the actuator signals, including alternate geometry or deviating process parameters, can change the actual design of the component and potentially lead to its failure during operation.
Yampolskiy will attempt to detect sabotage attacks by measuring signals to individual actuators.
Yampolskiy said the side-channel approach has three major advantages: it is non-invasive into the manufacturing process; is entirely interdependent of the original equipment manufacturer and the hardware, software and firmware of the additive manufacturing equipment; and will be hard to compromise because it can be operated from an air-gapped computer. In addition, it can be retro-fitted onto existing additive manufacturing machines.
“I have been working in this field since its inception and have closely followed several schools of thoughts on this problem,” Yampolskiy added. “This approach fits very well in my current understanding of how the threat of sabotage attacks should be addressed. At some point, when all goals of the project are achieved, perhaps it will help additive manufacturing industry to be safe from sabotage attacks.”
Media Contact: , jem0040@auburn.edu, 334.844.3447
