Hardware trojans detection based on circuit structure analysis and integrated model

Hardware security has become the focus of attention in today′s era, hardware Trojans detection techniques are also becoming more diverse. For the detection of hardware Trojans, an integrated model based on circuit structure analysis is proposed in this paper. In this method, three groups of gate level features of hardware Trojan were extracted by topological retrieval, and the complex circuit diagram was transformed into vector data format by NLP idea and method and Robust scaler standardized pretreatment. The hardware Trojan was detected by the integrated model of limit gradient lifting tree and long short memory neural network.The experimental results of 86.5% TPR and 97.9% TNR effectively illustrate the efficiency and accuracy of this method.