Electromagnetic Spectrum Analysis Based Hardware Trojan Detection Methodology

Abstract

Hardware Trojans can be implemented as modifications to ASICs, microprocessors, microcontrollers or digital-signal processors, and it has raised numerous concerns regarding possible threats to the basis of the network and communication systems. There have been a number of hardware Trojan detection methods, among which side-channel based method is one of the most popular. However, when analyzing the electromagnetic signals obtained from the chip, researchers tend to handle the signals in time domain, neglecting the ample information hidden in frequency domain. We propose a novel hardware Trojan detection methodology by analyzing the electromagnetic spectrum in frequency domain, and the experimental results show that this method is valid in identifying even very small Trojans.