Research On Hardware Trojan Detection Method Based On Logical Testing

Because of the globalization of the integrated circuit industry chain,the hardware Trojan has become an important factor affecting the information security of the integrated circuit,so the effective detection method of the hardware Trojan has become the current research hotspot.In this paper,to detect the hardware Trojan of combined activation and functional payload in the circuit,the hardware Trojan detection methods based on the combinatorial testing and the rare node are systematically studied respectively.The specific test vector generation algorithm is proposed,and the test vector generation scheme is designed respectively with the circuit.The main work and research creations of this paper are as the following:By analyzing the structure characteristics and classification methods of the hardware Trojan,the hardware Trojan of combined activation and functional load is determined as detection object in this paper.After analyzing the advantages,disadvantages and the applicability of several common hardware Trojan detection methods,this paper starts with the logical testing method,and analyzes the feasibility of detecting hardware Trojan based on the combinatorial testing and the rare node respectively,to improve the number and efficiency of activation of the hardware Trojan,the research direction of the test vector set generation method is studied.By analyzing the transmission relation of hardware Trojan attack signal in the circuit,a hardware Trojan detection model based on the combinatorial testing is established.With this model as the theoretical basis,to increase the number of activation of the hardware Trojan,a greedy strategy is designed to generate test vectors one by one,then a test vector generation algorithm based on greedy algorithm is proposed,and a test vector generation scheme is designed.To further improve the activation efficiency of hardware Trojan,the optimal expand algorithm is designed to generate test vectors by parameter expanding,then the test vector generation algorithm is proposed to improve the activation efficiency,and the test vector generation scheme is designed.To further study the method of detecting the hardware trojan based on rare nodes inside the circuit,the characteristics of rare nodes in the circuit are analyzed,the hardware Trojan logic detection model based on the rare node is established by analyzing the hardware Trojan attacks based on rare nodes.To enhance the pertinence of the test vector set and reduce the unnecessary test resource cost,the relationship between the number of the rare node activation times N and the number of hardware Trojan activation times S in the logical detection model is theoretically analyzed and deduced,and a reasonable method for determining the number of the rare node activation times N is put forward,then the test vector generation algorithm based on rare nodes is proposed,and the test vector generation scheme is designed.Experimental detection platforms based on the combinatorial testing and the rare node are built respectively,the AES-K2-i,AES-K3-i,AES-Trojan2,AES-Trojan3 and other test circuits are designed to simulate the attack of hardware Trojan designers,for each test circuit,simulation experiments are carried out respectively according to the proposed detection methods based on experimental detection platforms.The experimental results show that the proposed method based on greedy algorithm can can increase the number of activation of the hardware Trojan.The method of improving the activation efficiency can improve the number of activation of the hardware Trojan and the activation efficiency.The detection method based on the rare node can increase the number of activation of the hardware Trojan,improve the efficiency of detecting hardware Trojan,meanwhile,the targeting of test vectors is enhanced,which greatly reduces the cost of test resources.In conclusion,the detection method proposed in this paper can effectively improve the number of activation times and the activation efficiency,then enhance the success rate of detection hardware Trojan.