基于混合特征分析的硬件木马检测方法

Hardware Trojan detection method based on combined features analysis

提出一种基于混合特征分析的硬件木马检测方法,该方法首先在时序层级抽象并构建待测电路的控制数据流图,然后利用功能性分析方法建立以低动态翻转率为特征的动态可疑节点集,最终使用静态结构特征匹配方法实现硬件木马的检出.以Trust-Hub中涵盖Basic-RSA,AES和RS232基准电路在内的13种硬件木马为检测对象开展检测实验.实验结果表明:检测结果中硬件木马节点检出率达到100%,假阳性率控制在1.5%以内,检测准确率达到82.5%以上,证明了该检测方法的有效性.

A hardware Trojan detection method based on combined features analysis was proposed.The control data flow graph of the circuit under test was built first utilizing the gate-level netlist, and then nodes with low dynamic transition probability were screened to establish a suspicious set using functional analysis.The hardware Trojans could be finally detected by the corresponding static structural features.Experiments were carried out on Basic-RSA,AES and RS232 reference circuits in Trust-Hub covering 13 kinds of hardware Trojans. Experimental results show that the detection rate is 100% with a false positive rate under 1.5% and accuracy over 82.5%,demonstrating the effectiveness of the proposed method.