基于K-means聚类分析的硬件木马检测方法

Hardware Trojan Detection Based on K-means Algorithm

为检测芯片中存在的硬件木马,利用侧信道分析技术,采集功耗信息,并利用K-means算法对待测芯片进行聚类分析。利用K-means算法实现木马检测时,准确率与功耗数据中存在的噪声以及木马面积的大小有关,实验中植入了占芯片原始模块2.39%和1.49%的两种木马,均能够很好的检测出其中电路中的木马,验证了方法的有效性。

In order to detect the hardware Trojan in the integrated circuits, the K-means algorithm based on the side-channel analysis was used to classify the chip under tested. Detected the hardware Trojan by the K-means algorithm, the accuracy was related to the noise and the size of Trojan. Experimental results showed that the hardware Trojan occupied about 1.49% and 2.39% of the original module could be precise recognition, and verified the validity of the method.