This year, my research focused on two main topics. The first one is the research on modeling attack resilient, low power, and high throughput Strong PUF circuit that is used for IoT authentication, and the second one is the measurement of random telegraph noise in SRAM PUF after hot carrier injection (HCI) burn-in.

In terms of the Strong PUF research, I was involved in two projects. The first one is a 5-bit-output Strong PUF circuit with enhanced modeling attack resilience and response statistical quality. In previous research, the output response bit number was limited to 1 bit due to the non-ideal statistical quality and the vulnerability to modeling attacks. However, for authentication, a more extensive response bit number is required to ensure security. This work uses techniques such as XOR box, LUT blending, sub-challenge accumulation, and response accumulation. The 5-bit-output Strong PUF is realized with a better statistical quality and modeling attack resilience. The second Strong PUF project further pushes forward the cutting edge by achieving a 100-bit-output Strong PUF with close-to-ideal statistical quality and high modeling attack resilience. This is realized by implementing a complex permutation among the outputs of entropy sources and several techniques such as a challenge scheduler which masks the challenge inputs with the random entropy source data.

In the research of measurement of random telegraph noise (RTN) in SRAM PUF cells after HCI burn-in, a low-cost RTN observation method is developed. By applying external bias voltages, the mismatches of PUF bitcells can be measured repeatedly to observe how RTN changes them temporally. Also, this research measures and analyzes the relation between RTN amplitude and HCI burn-in time. It is observed that although HCI burn-in is an effective method to stabilize a PUF, it also enhances the amplitude of RTN, which might be harmful to a PUF’s stability even after stabilization and should be carefully considered when making a stabilization strategy.

As for publication, the Strong PUF research results in two international conferences and one domestic conference. The 5-bit Strong PUF circuit has been presented in IEEE ESSCIRC in Milan in September 2022. The security analysis of the 5-bit Strong PUF was also presented in SCIS (the biggest domestic security conference) in Kitakyushu in January 2023. The 100-bit Strong PUF will be presented at IEEE CICC in San Antonio in April 2023. The RTN research results in an international conference paper which has been presented in IEEE ICMTS in Tokyo in March 2023.