Themenbeschreibung: Spoofing is the transmission of fake Global Navigation Satellite System (GNSS) signals. It is a
malicious attack, which misleads a GNSS receiver to calculate the wrong position and time. A receiver,
which successfully detects a spoofing event, can warn a user that the position and time cannot be
trusted. Therefore, reliable spoofing detection is necessary for GNSS integrity.
Previous work on spoofing detection implemented rudimentary machine learning methods. It showed
that training on simulated data facilitates good performance when evaluating real recorded data.
However, there were several limitations in this study. First, the lack of variance in the real recorded data
was insufficient to allow successful training on the data and was adequate to evaluate the data. Second,
feature engineering was omitted, and leaves room for improvement. Furthermore, no sufficient studies
on the use of deep neural networks have been conducted.
This topic aims to improve spoofing detection by advanced machine learning techniques, including
feature optimization and deep learning methods. The evaluation is focused on generalization over
datasets, reliability and computational complexity.
Literature review on GNSS, GNSS spoofing, antenna arrays, feature engineering, and supervised machine learning
Analysis of provided datasets and identification of features
Identification of suitable machine learning methods (feature based or deep learning)
Implementation and optimization of the identified machine learning methods
Voraussetzungen: Basic knowledge of satellitenavigation and antenna theory/Knowledge of classification methodology and machine learning approaches/Familiarity with Python and machine learning frameworks such as tensorflow or scikit?learn
Themenbeschreibung: Low earth orbit (LEO) positioning, navigation and timing (PNT) has become a hot topic for
satellite?based navigation. LEO mega?constellations, i.e., constellations exceeding 1000 satellites, provide
excellent coverage and a good dilution of precision (DOP) for positioning. The much lower altitude – in
comparison to medium earth orbit (MEO) or geo?synchronous orbit (GSO) global navigation satellite
systems (GNSS) – also provides improved path?loss, yielding better signal to noise ratios (SNRs) for
detection, acquisition, and tracking. Furthermore, LEO satellites have high velocity; hence, allowing
Doppler?based positioning. Lastly, the LEO constellations selected for PNT are primarily communication
satellites, which makes them much less prone to spoofing attacks, in comparison to the predictable GNSS
satellites. In conclusion, LEO PNT is a good alternative or complimentary approach to legacy satellite
navigation.
LEO PNT does have several issues, as these signals are not designed for PNT. First, alternative acquisition
and tracking methods are required to extract appropriate observables from the satellites for navigation.
Second, these satellites do not necessarily provide their ephemeris information (orbital data) nor are
these precise, when they do. Therefore, accurate ephemeris information needs to be estimated. Third,
many LEO satellites do not necessarily have a timing reference; hence, this should also be determined.
Lastly, LEO satellites rarely have high precision atomic clocks like GNSSs, requiring additional clock
corrections. This project focuses on the first challenge, of designing and developing alternative
acquisition and tracking methods.
As an initial case study, Iridium is proposed. The Iridium downlink is in the 1.6 GHz frequency band,
which is adjacent to the L1 GNSS band. Therefore, state?of?the?art GNSS antennas, and modified GNSS
recording receivers may be used for a comparable evaluation. Although, Iridium is proposed, the student
may identify other suitable LEO constellations and adapt the project accordingly.
Literature review on GNSS, LEO-PNT methods, and acquisition-tracking methods
Analysis of opportunistic LEO constellations and selection of a suitable constellation
Signal recording and analysis of the selected LEO constellation
Identification of suitable acquisition-tracking architecture and extraction of observables
Implementation and of an acquisition-tracking architecture to show that a satellite can be used
Themengebiete: Global Navigation Satellite System, GNSS, Low Earth Orbit (LEO) Positioning, Navigation and Timing (PNT)
Voraussetzungen: Basic knowledge of satellite?navigation and tracking filters/Good knowledge on signal processing/Familiarity with Python and/or Matlab
Themenbeschreibung: Global Navigation Satellite Systems (GNSSs) provide global positioning capabilities to any
capable receiver — whether integrated into smartphones or used for aeronautical navigation. An issue with
GNSS signals are their low received signal power on earth’s surface, which makes them sensitive to
interference signals. Other licensed spectrum users, the result of old or poorly designed equipment leaking
in from adjacent frequency bands, opportunistic use of the spectrum by radio operators, or even
purposefully designed interferences, may be the cause interference signals. Therefore, it is crucial for the
local service quality to monitor the GNSS frequency bands and to report any interference signal to the
appropriate authorities.
The DARCY system consists of a network of low-cost sensor nodes to monitor the GNSS spectrum over a
larger geographic area. Each node consists of e.g. a Raspberry Pi Single Board Computer (SBC), a low-cost
GNSS receiver to monitor the GNSS signal quality, and a Software Defined Receiver (SDR) Frontend to
monitor the spectrum. The sensor nodes monitor and report any data to a server, which does interference
detection, interference localization, and reporting of any suspicious activity.
This master’s project focuses on algorithm development for collaborative interference localization. It uses
the DARCY sensor nodes to locate potential interference.
Literature review on GNSS, interference localization, Received Signal Strength (RSS) positioning, Time Difference of Arrival (TDoA) positioning, collaborative techniques, and information fusion
Analysis of interference localization methods and data available from the sensor nodes
Design and implementation of localization algorithm(s)
Setup of simulation environment to test selected algorithms
Field recording with interferences and sensor nodes
Themengebiete: Global Navigation Satellite System, Interference Localization, Received Signal Strength (RSS) Positioning
Voraussetzungen: Basic knowledge of satellite-navigation/Good knowledge of signal processing and estimation theory/Experience with Kalman Filtering is a bonus/Python and/or Matlab
Themenbeschreibung: The study of radar targets and their electromagnetic wave reflectivity models holds paramount importance in radar simulations. Traditionally, radar scenarios have operated under the assumption that the echo component is narrower than the radar resolution, allowing for the modeling of reflectivity as point scatter. However, with the current evolution of modern radar systems, such as Integrated Sensing and Communication (ISAC), i.e. for intelligent transportation and surveillance, this assumption no longer holds.
ISAC systems feature higher resolution in range, Doppler, and angle observations than the extension of echo-pattern, particularly in scenarios involving larger targets such as cars or unmanned aerial vehicles (UAVs). This necessitates the development of new reflectivity models capable of accurately capturing the characteristics of extended echo patterns, including consideration for micro-Doppler effects.
This thesis seeks to address this critical research gap by proposing novel reflectivity modeling approaches for extended echo patterns for radar targets. By investigating and developing these models, we aim to enhance the realism and accuracy of radar simulations in ISAC scenarios.
- Literature review on reflectivity models and radar cross section (RCS) for extended targets.
- Identify the analytical and computational requirements for reflectivity models of extended targets.
- Identification of suitable modeling approaches
- Implementation and development of new reflectivity models
- Evaluation of results
- Documentation of results and final thesis
The results achieved in the thesis are to be presented extensively in a written elaboration. The program source code must be sufficiently commented. The DFG rules for good scientific practice are to be respected in the preparation of the thesis. The execution of the work can be done in English as well as in German language.
Themengebiete:
Voraussetzungen:
Betreuer: Carsten Smeenk
Hochschullehrer: Prof. Albert Heuberger
Anfragen zu Themen und der Betreuung einer MA-Arbeit können sie auch unabhängig von den ausgeschriebenen Themen direkt an die Ansprechpartner der jeweiligen Forschungsschwerpunkte schicken.
