Research On Hardware Trojan Detection Technology Through Side Channel Signal Analysis Based On Improving Activation Probability

With the advent of the information age,integrated circuits?IC?have been applied to all walks of life.Wide range of integrated circuits applications occupies important positions even in key industry like military and national defense,so its security has been highly valued.Hardware Trojan is a special circuit with malicious purpose,which will be triggered under certain conditions.Once triggered,it'll tamper with the original circuit design or damage the circuit system,meanwhile it also can destroy the function of IC or leak confidential information,therefore Hardware Trojan brings great threat to the security of the IC.So Hardware Trojans and their detection technology have become the hot spot of current research in recent years.It is of great significance for both country and individuals to carry out research in this area.The article introduces the definition,insertion stage,damage and development status of hardware Trojans at the beginning.Then puts a research on the structure and principle of the hardware Trojans,classifies them according to different characteristics.In order to understand the hardware Trojans more deeply,three basic types of hardware Trojans are designed and verified by functional simulation.Finally,the current detection technology is summarized,and its characteristics are compared for comparison.Intensive studies are carried out mainly in the following three aspects:Firstly,we carry out a research on improving the activation probability of hardware Trojans.The logic cone structure of hardware Trojans trigger circuit is used to clarify the influence of the transition probability of the circuit node on the activation of the Trojan.Then we study the scheme of inserting the 2 to 1 Multiplexers?MUX?into the circuit node who has the low transition probability to improve the activation probability of the hardware Trojan,and analyze its principle and insertion process in detail.The simulation results show that the transition probability of the netlist nodes has been greatly improved after inserting MUXs by the algorithm,which is almost higher than the threshold(TP_th)we set.We further studied the extra overhead?power,area sand delay?causing by inserting MUXs,and control the increasing of the path delay within an acceptable range by setting the maximum delay ratio.Through our proposed method,we can increase the transition probability of entire circuit by inserting a small number of MUXs with low cost of path delay and area,thereby increasing the activation probability of the hardware Trojan.Second,we use side channel signals method based on power consumption to detect hardware Trojans.The principle of side channel signal detection hardware Trojan base on power is analyzed.Then we built a simulation experiment platform on the basis of inserting MUX to improve the activation probability of hardware Trojan.We present an algorithm on the basis of combining Principal Component Analysis?PCA?method with the Mahalanobis distance to reduce the dimension and optimize the power consumption data.The simulation results show that the insertion of MUXs can greatly promote the hardware Trojan detection based on power-side channel method.Moreover the PCA combining with the Mahalanobis distance can be very effective for dealing with the data,which can reduce the influence of redundant data and noise on detection,and greatly improve the success rate of detection.Finally,the hardware Trojans are detected using Multiple-parameters analysis technology,which can be viewed as an upgraded version of the side channel detection technology.After studying three methods of detecting hardware Trojans by exploring the correlation of two parameters which can show the characteristics of circuit,we use the multiple-parameters analysis method based on dynamic current-dynamic current(I_ddt-I_ddt)combining the scheme of improving the activation probability of hardware Trojan to detect hardware Trojans in different circuit.The results show that the improved hardware Trojan activation probability scheme can also promote the multiple-parameter analysis detection technology.And the multiple-parameter analysis technology has higher sensitivity for detecting hardware Trojans who has a larger area ratio,compared with the side channel signal detection technology based on power.