Visible to the public A General Framework of Hardware Trojan Detection: Two-Level Temperature Difference Based Thermal Map Analysis

TitleA General Framework of Hardware Trojan Detection: Two-Level Temperature Difference Based Thermal Map Analysis
Publication TypeConference Paper
Year of Publication2017
AuthorsShen, G., Tang, Y., Li, S., Chen, J., Yang, B.
Conference Name2017 11th IEEE International Conference on Anti-Counterfeiting, Security, and Identification (ASID)
ISBN Number978-1-5386-0533-2
Keywordscommercial chips, composability, cyber physical systems, differential temperature, differential thermal maps, equivalent circuits, field programmable gate arrays, general framework, Hardware, hardware Trojan detection, infrared imaging, integrated circuit design, Integrated circuit modeling, invasive software, Kalman filter, Kalman filter algorithm, Kalman filters, Mathematical model, normal circuits, operating chip, power proportion magnitude, pubcrawl, resilience, Resiliency, serious threats, Temperature measurement, Thermal analysis, Thermal map analysis, thermal map analysis detection method, trojan horse detection, Trojan horses, two-level temperature difference, White noise

With the globalization of integrated circuit design and manufacturing, Hardware Trojan have posed serious threats to the security of commercial chips. In this paper, we propose the framework of two-level temperature difference based thermal map analysis detection method. In our proposed method, thermal maps of an operating chip during a period are captured, and they are differentiated with the thermal maps of a golden model. Then every pixel's differential temperature of differential thermal maps is extracted and compared with other pixel's. To mitigate the Gaussian white noise and to differentiate the information of Hardware Trojan from the information of normal circuits, Kalman filter algorithm is involved. In our experiment, FPGAs configured with equivalent circuits are utilized to simulate the real chips to validate our proposed approach. The experimental result reveals that our proposed framework can detect Hardware Trojan whose power proportion magnitude is 10''3.

Citation Keyshen_general_2017