Researchers build novel malware detection system for Raspberry Pi
A research team has developed a malware detection system for Raspberry Pi. The group, based at France's Research Institute of Computer Science and Random Systems (IRISA), designed the anti-malware technique to scan the device's electromagnetic field.
A paper written by the team states that the system uses an oscilloscope paired with an H-field probe to detect abnormal electromagnetic behaviour, which would signal an attack. There is no additional software required for this method, which the group suggests makes this a novel approach.
The team uses side-channel information to "obtain precise knowledge about malware type and identity", so you know what threats are targeting the device. It can also work around tactics deployed by a bad actor to obscure the attack. What's more, it is hard for an attacker to detect the malware scanning system, given the lack of software.
The Raspberry Pi 2B used for this research was trained using a malicious data set and a safe baseline dataset. The group also evaluated their technique using Convolution Neural Networks (CNN).
The team concluded that the system they designed was 99.82% effective in its malware detection tests. If it was available for commercial use, the system could make it harder for devices to be hacked, as bad actors would also have to ensure electromagnetic field scans cannot detect their malicious code.
Source(s)
