Font Size: a A A

Hardware Trojan Detection Based On Gate-Level Circuit Structural Characteristics

Posted on:2019-09-27Degree:MasterType:Thesis
Country:ChinaCandidate:F Q ChenFull Text:PDF
GTID:2428330593951627Subject:Microelectronics and Solid State Electronics
Abstract/Summary:PDF Full Text Request
Development of IC results in fabless design and separation of design and fabrication,which in the same time relinquished the control of designers over the design and manufactory of IC.HT,namely Hardware Trojan,has emerged as a main concern of hardware security,and has posed potential trust and verification problems.Over the ten years or so since the concept of hardware Trojan is derived,HT detection methodologies has made great progress.Feature-based hardware Trojan detection methods are proposed,yet the feature database of HT need to be improved and expanded.In this paper a hardware Trojan detection method based on gate-level circuit structure characteristics is proposed.First,Trojan features are extracted from gate-level netlist of HT,and are expanded to a HT feature database.The HT feature database contains only four types of logic gates(AND,OR,NAND,NOR),which decreases the size of circuits to be analyzed.Then,the rare value algorithm is proposed to develop a measure of the rareness of a logic gate,which means the rareness of the output signal of this logic gate.Finally a rare value outlier determination algorithm is developed to flag the suspicious logic gates as Trojan candidates.We next improve the rare value algorithm to reduce detection complexity and runtime overhead,and improve the rare value definition to increase the accuracy rate of detection.As the experimental results show,with the proposed HT detection method using the improved rare value algorithm,the largest benchmark with 170 thousand units can be processed in 6 seconds.The proposed method can detect 39 out of 40 single triggered hardware Trojans from TrustHub website.
Keywords/Search Tags:hardware Trojan, gate-level circuit, structural characteristics
PDF Full Text Request
Related items