An innovative delay based algorithm to boost PUF security against machine learning attacks

摘要

Physical Unclonable Functions (PUFs) are the state-of-the-art topic in hardware oriented security and trust (HOST). PUFs have recently emerged as a promising solution for cyber security related issues including low-cost chip authentication, hacking of digital systems and tampering of physical systems. In this paper, a dynamic area efficient design along with an algorithm, namely, Optimal Time Delay Algorithm (OTPA) are proposed to improve the ROPUF challenge/response pairs (CRPs space). The generated CRPs space using this algorithm exhibit high PUF performance in terms of uniqueness and reliability. Experimental results demonstrate the competence of this innovative algorithm to enhance the CRP space using the probability of combinational statistic formula for improved security of the ROPUF against modeling attacks.