Lightweight PUF-Based Hardware Fingerprinting for Secure IoT Sensor Interfaces

Abstract

The rapid proliferation of Internet of Things devices has introduced unprecedented conveniences in industrial, medical, and consumer applications. However, the ubiquitous deployment of these resource-constrained devices has simultaneously expanded the attack surface, particularly at the physical sensor interface. Traditional cryptographic primitives often exceed the computational and power budgets of edge sensor nodes, necessitating alternative security paradigms. This paper proposes a lightweight, hardware-based fingerprinting mechanism leveraging Physical Unclonable Functions to secure Internet of Things sensor interfaces. By exploiting inherent manufacturing variations in standard complementary metal-oxide-semiconductor processes, Physical Unclonable Functions generate unique, unclonable digital signatures that serve as dynamic hardware fingerprints. We introduce a novel Ring Oscillator based Physical Unclonable Function architecture optimized for minimal area overhead and high reliability under varying environmental conditions. The proposed system integrates seamlessly into standard sensor communication buses, providing robust authentication and data integrity verification without the need for secure non-volatile memory key storage. Comprehensive experimental evaluations conducted on Field Programmable Gate Array platforms demonstrate that our design achieves near-ideal uniqueness and exceptional reliability across a wide range of temperatures and supply voltages. Furthermore, the lightweight nature of the fingerprinting protocol significantly reduces power consumption and latency compared to conventional cryptographic approaches, presenting a highly scalable solution for next-generation secure Internet of Things ecosystems.