首页|A strong physical unclonable function with machine learning immunity for Internet of Things application
A strong physical unclonable function with machine learning immunity for Internet of Things application
扫码查看
点击上方二维码区域，可以放大扫码查看
原文链接
万方数据
维普
The physical unclonable functions(PUFs)are novel cryptographic primitives in modern hardware security systems.Compared with traditional alternatives based on digital keys and non-volatile memory(NVM),the PUF system shows great unclonability,high efficiency,and physical attack resilience.However,the conventional PUF design suffers from weak machine learning immunity,high storage overhead,and unreliability,making it difficult to implement in the Internet of Things(IoT)and edge computing applications.This paper presents a new PUF design that could solve the proposed obstacles.By utilizing the emission probability of traps commonly found in nano-scaled transistors,a model-based PUF system with strong machine learning resistance could be achieved.This PUF design,called Prob-PUF,needs fewer challenge-response pairs(CRPs)space and reveals superior resistance to modeling attacks due to the mixture of stable/random bits in its output response.Moreover,the Prob-PUF system could reach a high level of uniqueness and robustness,making it a potential candidate for future cryptographically secured protocols within the IoT.
万方数据
维普
