Secure hardware key based on Physically Unclonable Functions and artificial Neural Network

Nowadays, Physically Unclonable Functions (PUFs) are widely used in hardware security applications such as device authentication, secret key generation and recently as a novel approach for securing Internet of Things (IoT) devices. Different methods have been proposed to create a unique binary string based on the PUF properties of silicon Integrated Circuit (IC). In this paper, the RO PUF and SR Latch PUF are implemented on a FPGA platform and their performance metrics such as reliability, uniqueness, and uniformity have been evaluated. An accurate and noiseless key is necessary for cryptography methods. To address this demand and increase reliability, we also propose two Error Correcting Code (ECC) methods based on Hopfield and Auto-associative memory Neural Networks. We show that the proposed error correcting code can successfully recognize and regenerate corrupted and noisy keys.