Spectral Singularities-Based Physically Unclonable Functions

Over the past few decades, the rapid growth of 5G and beyond technologies has substantially facilitated the advancement of the Internet-of-Things. However, the vast generation, exchange, and storage of information in this context may be exposed to significant threats associated with cyberattacks. Software-based cryptographic technologies, such as encrypted keys stored in local memory chips or cloud databases, have been the most widely used tools to counter such threats. Unfortunately, these tools have been proven vulnerable to machine learning-assisted modeling attacks. To overcome these limitations, various hardware-enabled encryption methods have emerged in recent years, such as physically unclonable functions (PUFs). In principle, the device-specified operation nature makes PUF a true random number generator, which can theoretically have near-ideal encryption quality. Thanks to the unclonable and unpredictable properties of PUF, it can effectively prevent invasive or modeling attacks and has become one of the most popular hardware secure alternatives. However, yet the most PUFs are silicon-based digital systems, which suffer from low device-to-device fluctuations and, therefore, low randomness and uniqueness, making them highly susceptible to artificial intelligent algorithms-related modeling attacks. To address these challenges, this dissertation proposes two classes of PUFs that respectively utilize two types of spectral singularities, divergent exceptional points, and coherent perfect absorber-laser points, existing in parity-time symmetric systems. We study their operation principles theoretically and verify their performances and feasibilities experimentally. The results demonstrate that both classes of radio frequency (RF) PUFs can achieve outstanding encryption quality compared to traditional silicon-based PUFs, with unprecedented resilience against machine learning-assisted attacks. We envision that our investigation may have a significant impact on the development of hardware-based cryptographic technology.