An intelligent classification method for Trojan detection based on side-channel analysis

Abstract

Side-channel analysis is an important strategy for Hardware Trojan (HT) detection. Karhunen-Love (K-L) expansion can be used to improve side-channel signals analysis quality, As an auxiliary post-processing method of K-L expansion, One Class Support Vector Machine (OCSVM) is introduced to achieve ICs intelligent classification. With the OCSVM and the power traces of Genuine ICs (Genuines), a hyper sphere can be built to distinguish the Trojan ICs (Trojans) from Genuines. The effectiveness of the proposed approach is experimentally demonstrated using power simulations performed on a representative circuit with several different Trojan circuits.