OTFS-Based Physical Layer Authentication for UAV-Assisted Data Collection in Wireless Sensor Network

Abstract

Autonomous aerial vehicle (commonly known as UAV) wireless communications have become widespread in military and civilian applications, partly due to its inherent line-of-sight (LoS) air-to-ground channels, easy deployment and high mobility. They have been adopted for data collection in the Internet of Things (IoT). For a communication scenario with a single UAV data collector in a delay-sensitive application, authentication is an important procedure which prevents illegal or malicious sensors from sending fake data to the data collector. However, traditional authentication schemes depend on cryptography, which is energy and time consuming. Thus, they are unsuitable for latency-intolerable applications and energy-constrained UAVs. This work aims to design a keyless, energy-efficient authentication scheme suitable for high-speed mobility scenarios. To this end, we propose an Orthogonal Time Frequency Space (OTFS)-based lightweight physical layer authentication (PLA) scheme. We begin by deriving the relationship between the location of the transmitter and normalized OTFS Doppler shift to develop the hypothesis test. Henceforth, we derive the probability density function (PDF) expressions of false alarm and missed detection. The PDF of the false alarm and missed detection are the basis for setting an optimal detection threshold for the authentication hypothesis test. Numerical results demonstrate that the proposed scheme holds well to fading effects for a robust and secure authentication scheme, and outperforms Orthogonal Frequency Division Multiplexing (OFDM) for PLA. The advantages of OTFS over OFDM for PLA are well elaborated.