Network-on-Chip hardware Trojan detection platform based on machine learning

Abstract

The extensive use of the Network-on-Chip (NoC) architecture makes it vulnerable to malicious attacks by hardware Trojans, especially Denial of Service (DoS) attack. To address this issue, this paper proposes a general NoC hardware Trojan detection platform based on machine learning. The platform establishes a security detection module including traffic feature tracking unit, feature registration unit, change point detection unit, and random forest detection unit, to accomplish the traffic-related hardware Trojan detection. The live-lock and fault routing Trojans are inserted in the proposed platform, then the simulation results verify the effectiveness of platform function and show its superiority to other existing detection schemes.