Machine-Learning-Based Multiple Abstraction-Level Detection of Hardware Trojan Inserted at Register-Transfer Level
Hau Sim Choo, Chia Yee Ooi, Michiko Inoue, Nordinah Binti Ismail, Mehrdad Moghbel, Sreedharan Baskara Dass, Chee Hoo Kok, Fawnizu Azmadi Hussin · 2019
Hardware Trojan refers to a malicious modification of an integrated circuit (IC). To eliminate the complications arising from designing an IC which includes a Trojan, it is suggested to apply Trojan detection as early as at register-transfer level (RTL). In this paper, we propose a hardware Trojan detection framework which consists of both RTL and gate-level classification using machine learning approaches to detect hardware Trojan inserted at RTL. In the experiment, all Trojan benchmarks were successfully identified without false positive detection on non-Trojan benchmark.