Dynamic Programming Based Grouping Method for RO-PUFs

Key generation is one of the most promising applications of Physical Unclonable Functions (PUFs), which requires 100% robust bit streams within each circuit and true randomness among a set of circuits. However, due to the noisy nature of PUFs, it is hard to provide stable outputs under changing environmental conditions, such as supply voltage and temperature. In this work, we have adapted Dynamic Programming (DP) to RO-PUFs for the first time in literature, in order to extract maximum entropy with minimum possible resource usage. Next, the robustness of all output bits is guaranteed even in unstable environmental conditions just by measuring a small subset of circuits prior to shipment. Finally, the efficiency of our method is analyzed and validated experimentally with FPGA implementation.