Graph‐based approach towards hardware Trojan vulnerability analysis

Abstract

The author proposes an incremental windowing technique based on graph technique to detect hardware Trojan triggering signal. First, they analyse the switching activity file generated during simulation. Then, based on the window size of choice, they extract paths of n-HT (n-hardware Trojan; n is the number of triggering signals) and compare the path information with the HT instances from HT database to determine HTs' earliest/farthest activation time. They experiment with the incremental windowing technique on 420 combinational HT instances for four combinational designs. The results indicate that small and stealthy HT is active for a limited set of the time window and the number of test patterns during these windows is compact instead of exhaustive traditional random vector based Automatic Test Pattern Generation (ATPG).