ANAMETRIC, INC — Department of Defense SBIR Phase II: AFX234-DCSO2

Description

The concept of identity is fundamental to cybersecurity. How do you know that a plane is friend or foe? Or that a remote device can be updated with new software, while keeping out updates from adversaries? How do you know a security badge is authentic? These functions and more are often enabled by a semiconductor circuit known as a PUF (Physical Unclonable Function) which is a digital fingerprint for chips.  This project intends to develop a new generation of chip-level secure ID capabilities with a new PUF technology. A PUF is a hardware structure with one or more unique manufacturing characteristics which can act as an immutable cryptographic root-of-trust (also known as a secure ID). Consequently, PUFs are an ideal foundation and are used in many secure applications, including supply-chain assurance/anti-counterfeiting, secure firmware updates, and trusted/secure communications - even ID badges. This project extends our previous work to build a novel hybrid quantum/classical photonics PUF. We will compare our theoretical model that was developed in Phase I against the operation of our first quantum hybrid PUF (QPUF) prototype chip. The company independently fabricated and characterized this first QPUF chip after Phase I ended. We will use the data from this analysis to drive further enhancements to this foundational technology. Recent applications of machine learning (ML) techniques have exposed flaws in current PUF designs, potentially compromising critical government and defense operational security – which can also jeopardize warfighter safety. We will conduct a cryptanalytic investigation of these types of ML-assisted threats against our QPUF to confirm its ability to withstand attacks of this nature.