Aging of SRAM PUFs: Mitigation and Advancements Through Machine Learning Techniques

Physical Unclonable Functions (PUFs) based on Static Random-Access Memory (SRAM) cells have gained significant attention for their role in hardware security. However, the aging of SRAM cells poses a critical challenge to the reliability and stability of SRAM PUF responses, potentially compromising their security properties. In recent years, machine learning techniques have emerged as promising solutions to mitigate aging effects in SRAM PUFs. This survey paper provides an analysis of the aging phenomena in SRAM PUFs and explores the application of machine learning algorithms for aging mitigation. We discuss the various aging mechanisms affecting SRAM cells and their impact on the reliability of SRAM PUF responses. Additionally, we review the state-of-the-art machine learning techniques employed for modeling and compensating aging effects in SRAM PUFs, including feature extraction, selection, model training, and optimization. Furthermore, we address the challenges and limitations associated with machine learning-based aging mitigation and propose future research directions. The findings of this survey highlight the potential of machine learning techniques in enhancing the reliability and security of aging-affected SRAM PUFs, opening new avenues for advancements in this field.