A system is developed that detects electromagnetic waves specific to malware and prevents attacks.
Aircraft and automobiles with automated driving carry the risk that 'a malicious change made by an attacker could lead to a serious accident.' A system has been developed that scans and detects such 'malicious changes' by scanning electromagnetic waves emitted from hardware.
System Sniffs Out Trojans in Electromagnetic Emissions
In January 2020, the Iranian military to shoot down civilian aircraft of Ukraine was mistaken for a military aircraft incident has occurred. IEEE Spectrum , a journal of electrical and information engineering, points out that malicious changes to automated weapons systems could lead to similar incidents. To prevent this situation, the US Defense Advanced Research Projects Agency (DARPA) has developed a Trojan horse attack prevention project 'SHEATH (Safeguards against Hidden Effects and Anomalous Trojans in Hardware)'. As a result, an AI system has been developed that scans the electromagnetic waves emitted by the hardware to detect malicious changes.
The following four images show the spectrum of electromagnetic waves emitted from the hardware to be detected, 'Clean Application' on the upper right is normal hardware, and 'V1 to V3' are hardware with malicious changes. Thing. The system developed this time has been learned to detect such 'electromagnetic waves peculiar to maliciously modified hardware', and its detection accuracy is said to be 98.87%.
Aether Argus, a company that participated in the development of the system, is working on the commercialization of the system. Also, according to Professor Angelos Keromytis of Georgia Institute of Technology, who also participated in the development, the modified detection capability of the system developed this time is useful not only for weapons but also for defense of 5G infrastructure.