Hardware Trojan Detection Technology Research Based On The HMM

With the globalization of integrated circuit design and manufacture,the safety of the integrated circuit products is more and more difficult to secure.Hardware Trojan is a new kind of threat.Hardware Trojan refers to the collective of redundancy circuits,that the attacker insert maliciously in the circuit at the processing of design and manufacture.It can make the circuit failure or disclose information in certain conditions.The presence of Hardware Trojan poses a serious threat to information security field,so the hardware Trojan detection technology has become the focus of the research topic both at home and abroad.In this paper,we carried out the related research work of hardware Trojan detection technology based on hidden markov models.Its main idea is to use the matching of side channel information that female parent circuit(incluing no hardware Trojan circuit)and contains hardware Trojan circuit to detect hardware Trojan.The main research content includes the following:Firstly,we had a detailed study of the characteristics and structure of hardware,which Laid a fundamental theoretical basis for the following hardware Trojan detection research work.Then,designed the hardware Trojan and hardware Trojan carrier circuit and simulated it,which provided possibility for subsequent testing.Secondly,set up a hardware Trojan detection platform,formulated the specific testing process,Carryied out experiments to collect data and ready for the data processing work.Finally,focused on the HMM algorithm to carry out the corresponding research work: first,in order to process,inductived hardware Trojan detection problem to the HMM specific problems of the three problems,Then extracted the characteristic parameters of female parent data and the Under test data;Used the HMM trained issue to trained the data of female,to get a model which represent the characteristics of the female parent data;Finally,used the evaluation problem of HMM to treat the data to be tested,obtained the matching degree of the under test data and the above model.after processed the data used hmm,again used a predetermined evaluation method to evaluate the test results.This thesis is successfully verified the effectiveness of HMM algorithm on the hardware Trojan detection by the experiment,at the same time in order to verify the applicability of this method,collected the electromagnetic radiation information again and used the above method processed,analied and verified.The results show that the hardware Trojan detection based on hidden markov model method can carry on the valid identification to the Trojan,and can detect the area ratio of 0.53% hardware Trojan.