A Hardware Trojan Detection Method for Gate-Level Netlists Employing the CAMELOT Measure

Nowadays, hardware designs are at an all-time high demand which has forced manufacturers to procure intellectual property from third-party vendors. Nevertheless, the third-party intellectual property can provide a route to major security concerns like hardware trojan. Most of the hardware trojan detection methods for such large circuits are deployed on side-channel analysis. The increase in process variations and the decrease in the size of the hardware trojan decreased the sensitivity of the side channel-based approach. In this proposed work, the testability feature of Computer-Aided MEasure for LOgic Testability, also known as CAMELOT is used to detect Hardware Trojans. This technique uses controllability - a combinational testability measure that is calculated using CAMELOT in detecting hardware trojans using the machine learning method. This work also makes use of the level of a logic gate as another feature in this detection process. This method deploys the K-Nearest Neighbour machine learning algorithm. The obtained results with experiments conducted on ISCAS-85 benchmark circuits demonstrate that CAMELOT, in addition to the level of the gate as features, detects hardware trojans with an accuracy and recall of 100%.