The Hardware Trojan Detection Based On Reference Curve

Integrated circuits are the core of computers and communication systems which deal with personal and government affairs,and it is closely related to our lives.Hardware Trojan is the malicious tampering of a circuit in the design or manufacturing phase of the integrated circuit,it aims to disclose critical information of the circuit or invalidate it under certain conditions.Because of the potential risks and hazards to modern information and communication systems brought by hardware Trojans,the hardware trojan detection technology has become a hot topic in the field of information security.According to the multi-parameter bypass analysis detection technology based on the reference curve,three methods based on the characteristics of the integrated circuits are proposed to reduce the influence of the process drift.They are the dynamic current and the maximum frequency(Iddt-Fmax),dynamic current and quiescent current(Iddt-Iddt0)and the dynamic current(Iddt1-Iddt0)under different test vectors.And these three methods are deduced theoretically.Then the EDA simulation platform and process are designed to verify the effectiveness of these three methods,The AES circuit is used as carrier circuit to verify the detection technology.In order to verify the environmental applicability of the detection technology,these three methods are simulated at different temperatures and voltages.Based on the preliminary simulation results,a hardware detection platform and testing flow based on FPGA are designed according to the measurement requirements.Different test schemes are designed for three detection methods.Based on the transient current analysis platform,the experimental verification of the three detection methods are completed.Finally,the test results are plotted and analyzed.The measured results show that the three methods have good detection effect.The type of Trojan horse is the timing "time bomb",the size of the horse is 4.54%.In order to further improve the detection coverage of hardware trojans,this paper presents two solutions,the first one is to partition the design and insert the scan chain in it,the impact of trojan circuit on the bypass parameter information will be manifested by measuring the different scan chain.And choosing the CAN bus controller as the experimental circuit,partition the circuit into three parts and insert the scan chain in each part.After completing the tape-out,a hardware detection platform is set up,and the multi-parameter bypass analysis technology based on the reference curve is used to detect the hardware trojan.The second solution is to add the same test vector to the AES circuit to observe the current value in different timing windows.If the change of the bypass information value in these timing windows is larger than the current value,it can be determined that the circuit contains hardware Trojan.This method is also called self-comparison technology.Respectively,these two methods are analyzed in detail,and in order to achieve automated test,an automated test software is designed for this method.The final test results show that the first method can improve the detection resolution from 1.42% to 2.64%.The second method can detect the hardware Trojans hidden in the third party IP when the reference curve can't be obtained,and it can improve the hardware Trojan detection coverage effectively.