FLAM-PUF: A Response–Feedback-Based Lightweight Anti-Machine-Learning-Attack PUF

Wu, Linjun; Hu, Yupeng; Zhang, Kehuan; Li, Wenjia; Xu, Xiaolin; Chang, Wanli

doi:10.1109/tcad.2022.3197696

Cited by 16 publications

(7 citation statements)

References 51 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Meanwhile, we also simulated the avalanche characteristics of mainstream PUFs. Apart from APUF, the resource consumption of these PUF structures is similar to our proposed structure [ 25 ]. As shown in Figure 14 , these PUFs do not exhibit ideal statistical characteristics.…”

Section: Numerical Experimentsmentioning

confidence: 88%

“…According to the results of Shah et al [ 32 ], when the number of feedback iterations increases to 8, the reliability of the response output is less than 65%. Similarly, FLAM-PUF [ 25 ] introduces an LFSR between the challenge and the PUF and changes the feedback coefficients of the LFSR through two confusion stages, introducing nonlinearity and randomness. In the first stage, a single-bit response is fed back to a certain feedback coefficient of the LFSR, undergoing n − 1 cycles of confusion and collecting n − 1 bits of response.…”

Section: Related Workmentioning

confidence: 99%

“…Unfortunately, besides LSPUF and SPUF, MPUF [ 14 ] and [ 24 ] have only discussed SAC. FLAM-PUF [ 25 ] introduces response cyclic feedback, greatly enhancing the GSAC of PUF, but due to the influence of environmental noise, it is difficult for this structure to maintain good reliability, and the LFSR feedback polynomial after secondary confusion may not necessarily be the primitive polynomial, which will reduce the CRPs space. In response to the aforementioned security and reliability issues, this article mainly makes the following contributions: It proposes a response feedback PUF based on LFSR and APUF, which greatly enhances the nonlinearity and randomness of the structure and enhances its ability to resist ML modeling and GSAC through response cyclic feedback.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

A Response-Feedback-Based Strong PUF with Improved Strict Avalanche Criterion and Reliability

Zhu,

Jiang,

Huang

et al. 2023

Sensors

View full text Add to dashboard Cite

Physical Unclonable Functions (PUFs) are significant in building lightweight Internet of Things (IoT) authentication protocols. However, PUFs are susceptible to attacks such as Machine-Learning(ML) modeling and statistical attacks. Researchers have conducted extensive research on the security of PUFs; however, existing PUFs do not always possess good statistical characteristics and few of them can achieve a balance between security and reliability. This article proposes a strong response-feedback PUF based on the Linear Feedback Shift Register (LFSR) and the Arbiter PUF (APUF). This structure not only resists existing ML modeling attacks but also exhibits good Strict Avalanche Criterion (SAC) and Generalized Strict Avalanche Criterion (GSAC). Additionally, we introduce a Two-Level Reliability Improvement (TLRI) method that achieves 95% reliability with less than 35% of the voting times and single-response generation cycles compared to the traditional pure majority voting method.

show abstract

Section: Numerical Experimentsmentioning

confidence: 88%

Section: Related Workmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

A Response-Feedback-Based Strong PUF with Improved Strict Avalanche Criterion and Reliability

Zhu,

Jiang,

Huang

et al. 2023

Sensors

View full text Add to dashboard Cite

show abstract

“…The proposed PUF fingerprint offers an additional benefit over challenge-response-based PUFs. With the advancement of machine learning techniques, it has become increasingly challenging to protect challenge-response PUFs against machine learning cloning attacks [ 67 , 68 ]. This provides an additional reason to develop fingerprint PUFs, which are used locally and, as a result, are less susceptible to machine learning attacks.…”

Section: Measurementsmentioning

confidence: 99%

“…The second method requires a third party that stores valid challenge-response pairs [ 66 ]. While challenge-response PUFs are considered strong PUFs, they are vulnerable to Machine Learning attacks [ 67 ], which can model PUF behavior based on collected challenge-response pairs [ 68 ].…”

Section: Introductionmentioning

confidence: 99%

Microcontroller-Based PUF for Identity Authentication and Tamper Resistance of Blockchain-Compliant IoT Devices

Vinko,

Miličević,

Lukić

et al. 2023

Sensors

View full text Add to dashboard Cite

Blockchain-based applications necessitate the authentication of connected devices if they are employed as blockchain oracles. Alongside identity authentication, it is crucial to ensure resistance against tampering, including safeguarding against unauthorized alterations and protection against device counterfeiting or cloning. However, attaining these functionalities becomes more challenging when dealing with resource-constrained devices like low-cost IoT devices. The resources of IoT devices depend on the capabilities of the microcontroller they are built around. Low-cost devices utilize microcontrollers with limited computational power, small memory capacity, and lack advanced features such as a dedicated secure cryptographic chip. This paper proposes a method employing a Physical Unclonable Function (PUF) to authenticate identity and tamper resistance in IoT devices. The suggested PUF relies on a microcontroller’s internal pull-up resistor values and, in conjunction with the microcontroller’s built-in analog comparator, can also be utilized for device self-checking. A main contribution of this paper is the proposed PUF method which calculates the PUF value as the average value of many single PUF measurements, resulting in a significant increase in accuracy. The proposed PUF has been implemented successfully in a low-cost microcontroller device. Test results demonstrate that the device, specifically the microcontroller chip, can be identified with high accuracy (99.98%), and the proposed PUF method exhibits resistance against probing attempts.

show abstract

ML for Attack and Defense of PUFs: Current Status and Future Directions

Ferens,

Dushku,

Kosta

2023

Lecture Notes in Networks and Systems

View full text Add to dashboard Cite

FLAM-PUF: A Response–Feedback-Based Lightweight Anti-Machine-Learning-Attack PUF

Cited by 16 publications

References 51 publications

A Response-Feedback-Based Strong PUF with Improved Strict Avalanche Criterion and Reliability

A Response-Feedback-Based Strong PUF with Improved Strict Avalanche Criterion and Reliability

Microcontroller-Based PUF for Identity Authentication and Tamper Resistance of Blockchain-Compliant IoT Devices

ML for Attack and Defense of PUFs: Current Status and Future Directions

Contact Info

Product

Resources

About