Detecting AIS position spoofing using LEO satellites, Doppler shifts, and MCMC methods.

Gregersen, Emil Skovfoged ; Nissen, Martin Mølbach

4. semester

Mathematical Engineering, Master

2018

2018-06-01

199 pages

This master thesis is written in conjunction with nano-satellite manufacturer GomSpace. It concerns the use of Doppler shift data extracted from AIS signals received by low-earth-orbit satellites to detect AIS position spoofing. Vessels exceeding a certain size are required by law to send out AIS signals with their position encoded. The problem is that a vessel can falsify this information in said AIS signals. The act of doing so is referred to as AIS position spoofing. Spoofing of a vessel’s position can be of interest if said vessel is e.g. fishing in protected waters. The system encompassing the vessel, the radio channel and the satellites is referred to as the space-based AIS system. The observable and unobserved variables that can aid in AIS position spoofing detection in this system are explored in this thesis. In order to determine whether AIS position spoofing is occuring in a given scenario, probability models describing the relationships between the variables in the system are created. The spoofing detection then consists of performing inference in the probability models using Markov chain Monte Carlo methods. The scenarios in which the developed methods are tested are constructed in order to determine how far away from its true position a vessel has to spoof its position in order for the developed methods to detect the spoofing. Results show that in some cases the developed methods can detect spoofing when a vessel is spoofing its position 5-10 km away.

Satellit ; Markkæde Monte Carlo ; Von Mises ; Doppler skift ; AIS ; Position spoofing ; MCMC ; Gibbs sampler ; Metropolis Hastings

Download
View record in AAU Student Projects