Researchers at Rutgers University in New Brunswick and Georgia Institute of Technology have devised three techniques to determine if 3D printers have been hacked and pose any dangers to a facility. Their findings were compiled into a study published at the 26th USENIX Security Symposium in Vancouver, British Columbia. “With the advent of 3-D bioprinting, there needs to be an increased effort in the verification of 3-D printed organic material, such as organs or tissues,” said Luis Garcia, one of the authors of the study. “Our three methods of verification will be able to detect any imperfections in the 3-D printed model that were introduced either maliciously or by user error. If we can detect an imperfection early in the printing process, we can shut down the printing and save both time and materials.”
The first two techniques consisted of studying the movement of printer parts plus the sounds made by the printer. Filament, which consists of materials such as plastic or metal, was pushed through the extruder (also known as the “arm” of the printer) to form objects. During this process, researchers used sensors to track the movements of the extruder, and microphones to monitor the noises made during printing. The third technique involved using CT scanners and Raman spectroscopy to examine printed objects. For this test, tiny gold nanoparticles were injected into the filament to act as contrast agents. The filament was inserted through the extruder to produce objects, which were then scanned to study shifts in the position of the nanoparticles, as well as where there may be holes and other defects.
Though the CT and Raman spectroscopy tests showed some progress in detecting defects, it doesn’t necessarily mean that they’re the most effective scanning tools in every instance. “For the effectiveness of these scanning modalities, it really is domain-specific,” Garcia added. “The choice of the scanner would depend on the materials required for the print as well as the depth of the secondary implanted material. For instance, Raman spectroscopy has depth limitations, but is effective at detecting particular materials. The CT scanner provides a greater depth of scanning, but this requires the embedded markers to have a high contrast in X-ray density from the base material.”
The authors of the study believe that though these methods will not become standard forms of protection against cyberattacks, but they will act as additional tools to assist domain-specific verification methods. “These specific methods will probably not be standard tools, but we do envision the overall concept of using complementary verification methods to become standardized for both run-time and post-production verification, in order to ensure integrity for safety-critical applications, as well as to save on costs of time and materials,” Garcia said. The authors will continue to search for more ways to combat cyberattacks in 3 D printers and propose further forms of defense.