Gate-level Hardware Trojan Detection Method Based On Support Vector Machine

In recent years,with the rapid development of the integrated circuit industry,transistor technology has continued to shrink,and the scale of integrated circuits has continued to expand,resulting in rising manufacturing costs and manufacturing cycles.In order to reduce manufacturing costs and manufacturing cycles,most integration companies use third-party EDA tools and IP cores in the design stage,and choose to outsource manufacturing processes to third-party foundries during the manufacturing stage.Due to the intervention of these third parties,it is possible to implant circuits with malicious functions in the chip.And this kind of circuit with malicious function is the hardware Trojan that this article will study.These untrusted third parties are the main source of hardware Trojan,and hardware Trojan is currently the biggest threat to integrated circuits.How to prevent the implantation of hardware Trojan or to detect hardware Trojan has become a research hotspot in recent years.The research focus of this article is the detection strategy of the hardware Trojan.Existing more successful hardware Trojan detection methods are mainly based on side-channel analysis and logic-testing methods.They do have a good detection effect,but they are often faced with difficulties such as hardware Trojan activation difficulty.In recent years,many scholars at home and abroad have successfully applied the big data analysis capabilities of machine learning algorithms to the detection of hardware Trojan,and they have achieved very good results,but there is still a problem of low detection rate.In order to deal with the above problems,this paper proposes a gate-level hardware Trojan detection method based on Support Vector Machine(SVM).The main research contents and contributions are as follows:(1)In order to deal with the problem of low detection rate of hardware Trojans,this paper analyzes the unique circuit structure,functions and insertion methods of hardware Trojans,extracts the features related to hardware Trojans,and successfully uses the Support Vector Machine algorithm to realize the detection of hardware Trojan.The experimental results of the reference circuits on Trust-Hub show that the method in this paper can achieve an average hardware Trojan detection rate of up to 93%,and the detection rate of some reference circuits reaches 100%.Compared with the existing detection methods based on machine learning,this paper has greatly improved the detection rate.(2)In view of the existing detection methods based on side-channel analysis and logic testing,facing the problem of needing to activate the hardware Trojan in advance,this paper designs three common hardware Trojans and implants them in the low controllability signal inside the reference circuit,making its activation probability is very low.The experimental results of 15 circuits show that the method based on Support Vector Machine can achieve good detection effect,and has the advantage of not requiring the test vector to activate the hardware Trojan,effectively avoiding the hardware Trojan activation problem.