A region based fingerprinting for hardware Trojan detection and diagnosis
T. K. Saran, Sree Ranjani Rajendran, M. Nirmala Devi · 2017
In today's scenario, any design is vulnerable to third party attackers like Hardware Trojan depends on their accessibility to the design. Hardware Trojans are the pernicious circuitry which are added to an original design by an attacker. These Trojans may lead to erroneous output of a system or even can incapacitate the design. Hence there is a demand for detecting the Trojans. The proposed method segments the whole design into small regions and generates fingerprint out of those. The number of fingerprints again reduced by selection of only regions containing low activity nodes in the first iteration due to the probability of inserting a Trojan in the low activity node is more. Segmentation helps to improve to locate where exactly the Trojan is inserted. This proposed method is suitable to detect Trojans which gets activated by aging or Trojan that gets activated after a specific time period. Hamming distance Pattern optimization technique improves the toggling inside the regions which makes the Trojan effect more observable. Proposed algorithm is implemented and validated using ISCAS'85 benchmark circuits.