Physical layer security for IoT applications

The increasing demands for Internet of things (IoT) applications and the tremendous increase in the volume of IoT generated data bring novel challenges for the fifth generation (5G) network. Verticals such as e-Health, vehicle to everything (V2X) and unmanned aerial vehicles (UAVs) require solutions that can guarantee low latency, energy efficiency,massive connectivity, and high reliability. In particular, finding strong security mechanisms that satisfy the above is of central importance for bringing the IoT to life. In this regards, employing physical layer security (PLS) methods could be greatly beneficial for IoT networks. While current security solutions rely on computational complexity, PLS is based on information theoretic proofs. By removing the need for computational power, PLS is ideally suited for resource constrained devices. In detail, PLS can ensure security using the inherit randomness already present in the physical channel. Promising schemes from the physical layer include physical unclonable functions (PUFs), which are seen as the hardware fingerprint of a device, and secret key generation (SKG) from wireless fading coefficients, which provide the wireless fingerprint of the communication channel between devices. The present thesis develops several PLS-based techniques that pave the way for a new breed of latency-aware, lightweight, security protocols. In particular, the work proposes: i) a fast multi-factor authentication solution with verified security properties based on PUFs, proximity detection and SKG; ii) an authenticated encryption SKG approach that interweaves data transmission and key generation; and, iii) a set of countermeasures to man-in-the-middle and jamming attacks. Overall, PLS solutions show promising performance, especially in the context of IoT applications, therefore, the advances in this thesis should be considered for beyond-5G networks.

Name dc:type.qualificationname

phd

Level dc:type.qualificationlevel

doctoral

Grantor dc:publisher.institution

University of Essex

Year dc:date.issued

2020

Author dc:creator

• Mitev, Miroslav

Language dc:language

en