Ph.d.-forsvar Søren Reime Larsen

Principal supervisor:

Associate professor Daniel Haugård Olesen, DTU

Co-supervisors:

Anna B.O. Jensen, AJ Geomatics, Denmark
Lars Stenseng, Bedrock Solutions, Denmark

Examiners:

Professor Jørgen Dall, Space DTU
Professor Heidi Kuusniemi, Universithy of Vaasa, Finland
Senior Researcher Aiden James Morrison, SINTEF, Norway

Chairperson at defence:

Professor Per Knudsen

Summary:

This dissertation investigates the possibilities of detecting and locating sources of radio-frequency interference (RFI) in the GNSS satellite bands, with a primary focus on the L1-band at 1575.42 MHz.

The objective is to use commercial, mass-market GNSS receivers as sensors for both detection and localization of GNSS interference. In the new approach suggested here, the carrier-to-noise ratio (C/N0), is used to estimate the distance to the source of interference – in this case a GNSS jammer. Having three or more estimated distances the source can be located using a tri- or multi-lateration approach, commonly described as the method of intersecting circles.

C/N0 is a common measure of signal strength reported by most modern GNSS receivers. To understand how it is useful for localization of a GNSS jammer, this thesis first covers how C/N0 is computed, how a jamming signal propagates, and how it affects the C/N0 observations of a GNSS receiver.

Using this knowledge a jamming detection algorithm for C/N0 is implemented and used on data from geodetic reference stations to investigate the amount and extent of GNSS interference in Denmark. It is shown that GNSS interference is a common event, but these events are often of short duration and low-powered. Interference is also mostly registered near major roads, suggesting the sources are found in cars passing by the receivers.

The impact of jamming on GNSS precise positioning is then evaluated using the TAPAS network of GNSS reference stations in the Danish city of Aarhus. This investigation is based on actual jamming found in the TAPAS stations, and shows that in most cases of real-world jamming, using all the available GNSS signals, largely mitigates the impact of the interference.
Finally, using data from jamming tests performed in Norway, this work investigates the accuracy with which a real-world GNSS jammer can be located using the new approach. The localization accuracy is evaluated using a setup of various modern Android smartphones, which has not previously been attempted outside of a lab. Furthermore, the accuracy is also evaluated for a solution using identical survey-grade receivers.

It is shown that a jammer can be located, but that there are several disadvantages to the approach. Using real-world data, it is seen, that the localization model is inconsistent, especially for smartphones, with low-grade GNSS hardware. Therefore, a novel observation-exclusion algorithm is implemented, and it is shown that it significantly increases the accuracy of the solution.