Decentralized Detection of GPS Spoofing in Vehicular Ad Hoc Networks

This letter proposes a decentralized scheme for the detection of global positioning system (GPS) spoofing in vehicular ad hoc networks. In this scheme, vehicles exchange their measured GPS code pseudo-ranges with neighboring vehicles using dedicated short-range communications. The vehicles then perform linear operations on the exchanged GPS data and derive independent statistics that are related to the measurements of each neighbor. Using these statistics, a vehicle implements a cumulative sum procedure to locally detect high correlations in the time of arrival of spoofed GPS signals. The vehicles report their local detection values to an elected head vehicle, which employs a min-max change-detection procedure to optimize the global-spoofing detection. Our experimental results show that: 1) the average detection delay decreases as the number of vehicles increase and 2) the proposed scheme converges to optimal as the probability of false-alarm reduces to zero.

I. Introduction

Vehicular ad hoc networks (VANETs) are decentralized networks that are formed spontaneously when intelligent vehicles come in range to one another. These networks facilitate wireless vehicle-to-vehicle (V2V) and vehicle-to-infrastructure (V2I) connectivity through dedicated short-range communications (DSRC). Equipped with on-board-units (OBUs), vehicles can exchange their state information (e.g. position) with neighboring vehicles, and nearby roadside units (RSUs) in a VANET, enabling intelligent applications to improve transportation safety and reduce traffic emissions [1].