Open theses and project topics
Details about the currently avaliable topics can be found on SiROP. The categories (Bachelor Thesis, Master Thesis and Semester Project) are primarily recommendations. Depending on the topic, it might be possible to adapt the scope to make it suitable for a different type of thesis/project.
Detection of abnormal GNSS ephemerides with machine learning
A machine learning-based classifier will be developed to automatically detect abnormal GNSS ephemerides in daily broadcast ephemeris files.
Keywords
GNSS ephemeris, Machine learning, classification
Labels
Semester Project , ETH Zurich (ETHZ)
Description
Goal
Contact Details
More information
Published since: 2024-05-06 , Earliest start: 2024-09-16 , Latest end: 2024-12-20
Applications limited to ETH Zurich
Organization Space Geodesy (Prof. Soja)
Hosts Soja Benedikt
Topics Earth Sciences
Implementing a dynamic noise model to improve drone-based GNSS RTK tropospheric estimation
A dynamic tropospheric process noise model will be implemented into GNSS real time kinematic (RTK) algorithms to improve the estimation of drone-based GNSS zenith total delays (ZTDs).
Keywords
GNSS, drone, tropospheric process noise, atmospheric monitoring
Labels
Semester Project , ETH Zurich (ETHZ)
Description
Goal
Contact Details
More information
Published since: 2024-05-06 , Earliest start: 2024-09-16 , Latest end: 2024-12-20
Applications limited to ETH Zurich
Organization Space Geodesy (Prof. Soja)
Hosts Soja Benedikt
Topics Earth Sciences
Gaussian Processes for Ionospheric Modelling
This thesis explores ionospheric modeling using GNSS and potentially VLBI data, employing Gaussian Process regression to address the non-linear behaviors and noise inherent in such data. The study focuses on enhancing predictive accuracy and the quantification of uncertainties in ionospheric variations, which are essential for improving global navigation and communication systems.
Keywords
Ionosphere, GNSS, VLBI, Gaussian Processes, machine learning
Labels
Bachelor Thesis , ETH Zurich (ETHZ)
Description
Goal
Contact Details
More information
Published since: 2024-05-06 , Earliest start: 2024-09-16 , Latest end: 2024-12-20
Applications limited to ETH Zurich
Organization Space Geodesy (Prof. Soja)
Hosts Soja Benedikt
Topics Earth Sciences
Implement better vegetation model for the GNSS-IR soil moisture retrieval
This master thesis aims to improve the retrieval of soil moisture using the GNSS Interferometric Reflectometry (GNSS-IR) method by the development of a new, machine-learning based model for the correction the vegetation influence on the retrieval.
Keywords
GNSS Interferometric Reflectometry, Vegetation, Machine-learning
Labels
Master Thesis , ETH Zurich (ETHZ)
Description
Goal
Contact Details
More information
Published since: 2024-05-06 , Earliest start: 2024-08-01 , Latest end: 2025-05-31
Applications limited to ETH Zurich
Organization Space Geodesy (Prof. Soja)
Hosts Soja Benedikt
Topics Earth Sciences
Physics-Informed Neural Networks for Regional Geoid Modeling
This master thesis aims to explore the application of Physics-Informed Neural Networks (PINNs) to regional geoid modeling. PINNs integrate physical constraints into neural network architectures, offering a novel approach to accurate geoid modeling while maintaining interpretability.
Keywords
Regional geoid modeling, Gravity field, Physics-Informed Neural Network (PINN), Machine learning
Labels
Master Thesis , ETH Zurich (ETHZ)
Description
Goal
Contact Details
More information
Published since: 2024-05-06 , Earliest start: 2024-08-01 , Latest end: 2025-05-31
Applications limited to ETH Zurich
Organization Space Geodesy (Prof. Soja)
Hosts Soja Benedikt
Topics Earth Sciences
Detecting water bodies from SWOT measurements using deep learning
In this study, the student should apply deep learning algorithms to segment the measurements from the Surface Water and Ocean Topography satellite mission, specifically focusing on inland water bodies. The outputs of this study may contribute to inland water detection during flood events and also potentially to refining the pre-defined water body shapes.
Keywords
Satellite altimetry, SWOT, Segmentation, inland water
Labels
Master Thesis , ETH Zurich (ETHZ)
Description
Goal
Contact Details
More information
Published since: 2024-05-06 , Earliest start: 2024-08-01 , Latest end: 2025-05-31
Applications limited to ETH Zurich
Organization Space Geodesy (Prof. Soja)
Hosts Soja Benedikt
Topics Earth Sciences
Quantifying contributions of individual water storage components to terrestrial water storage using a generative model
In this study, the student should develop a generative model to separate the total signals measured by GRACE(-FO) satellite missions into the contributions of individual water storage components. The results will be evaluated by comparing them with independent in-situ and satellite-based storage observations. The findings of this study will contribute to a better understanding of the terrestrial water cycle from a global perspective.
Keywords
Terrestrial water storage, GRACE(-FO), Hydrological model, Generative model
Labels
Semester Project , ETH Zurich (ETHZ)
Description
Goal
Contact Details
More information
Published since: 2024-05-06 , Earliest start: 2024-09-16 , Latest end: 2024-12-20
Applications limited to ETH Zurich
Organization Space Geodesy (Prof. Soja)
Hosts Soja Benedikt
Topics Earth Sciences
Quasi-Continuous VLBI observations
This work will explore the potential of greatly increasing the number of VGOS sessions per month by limiting the amount of recorded data.
Keywords
VLBI, VGOS
Labels
Semester Project , ETH Zurich (ETHZ)
Description
Goal
Contact Details
More information
Published since: 2024-05-06 , Earliest start: 2024-09-16 , Latest end: 2024-12-20
Applications limited to ETH Zurich
Organization Space Geodesy (Prof. Soja)
Hosts Soja Benedikt
Topics Earth Sciences
Total Electron Content Derived from Smartphone GNSS Data for Ionosphere Monitoring
Global Navigation Satellite System (GNSS) is a well-recognized technique for ionosphere monitoring as additional signal contributions, that GNSS signals experience while traversing the ionosphere, are proportional to the total electron content (TEC). Currently, the operational and publicly available GNSS-based global ionosphere maps are exclusively derived from slant TEC (STEC) acquired using geodetic-grade receivers. GNSS observations that are potentially available at large scale from affordable smart devices, such as smartphones, could form an attractive source of information to improve the spatio-temporal resolution of GNSS data sets that are available for the analysis in this subject. Since 2016, Android phone users can access the raw GNSS data within smartphones. In addition, modern smartphones can track nowadays dual-frequency multi-GNSS signals, which makes extraction of carrier-phase-based VTEC possible. The Space Geodesy group at ETH Zurich initialized a crowdsourcing campaign in March 2022 for collecting smartphone observations from interested participants. Up to now, several TBs of smartphone GNSS data has been collected globally, which provides a great opportunity to study the potential of ionosphere modelling using smartphones.
Keywords
Toal Electron Content, smartphone, GNSS, ionosphere monitoring
Labels
Master Thesis , ETH Zurich (ETHZ)
PLEASE LOG IN TO SEE DESCRIPTION
More information
Published since: 2024-05-02 , Earliest start: 2024-08-01 , Latest end: 2025-05-31
Applications limited to ETH Zurich
Organization Space Geodesy (Prof. Soja)
Hosts Soja Benedikt
Topics Engineering and Technology , Earth Sciences
Impact of satellite mega-constellations on VLBI observations
Very Long Baseline Interferometry (VLBI) is versatile space geodetic technique, measuring radiation from extragalactic radio sources. VLBI supports a wide range of applications, including the determination of the terrestrial and celestial reference frames, measurements of the full set of Earth orientation parameters, and the determination of geophysical models. Due to its observation principle, VLBI observations have to be actively planned and organized between the radio telescopes. In recent years, a dramatic growth in the number of satellites emitting radio frequencies can be seen. Most notably is the Starlink system, with more than five thousand satellites being in orbit already. These satellite mega-constellations threaten VLBI observations since their emitted artificial radio signals interferes with the natural radio signal emitted from the extragalactic radio sources, especially if the satellite is aligned with the observed radio source. So far, no active measures are in place to avoid VLBI observations in such cases.
Keywords
VLBI, satellite mega-constellations, Starlink, satellite orbit
Labels
Master Thesis , ETH Zurich (ETHZ)
PLEASE LOG IN TO SEE DESCRIPTION
More information
Published since: 2024-05-02 , Earliest start: 2024-08-01 , Latest end: 2025-05-31
Applications limited to ETH Zurich
Organization Space Geodesy (Prof. Soja)
Hosts Soja Benedikt
Topics Engineering and Technology , Earth Sciences
MPGNET: Feasibility assessment for real-time GNSS Meteorology
The chair of Space Geodesy at ETH Zürich is currently in the process of building up a multi-purpose network of low-cost GNSS payloads (MPGNET), collocated at different meteorological observation sites in Switzerland, operated by MeteoSwiss. These collocated GNSS/Meteo sites are expected to provide benefits for both the meteorological and geodetic community. Applications are manifold and reach from improved GNSS troposphere products to monitoring of soil moisture and snow heights in the vicinity of the station. First tests have already shown that the data quality delivered by the payload is high enough to produce high-quality troposphere products, even in real-time. This thesis further develops the idea of real-time streaming and processing of GNSS raw data for meteorological applications. This should be done in three steps: 1) Installation of a Networked Transport of RTCM via Internet Protocol (Ntrip) server/client at the prototype station in Zurich-Affoltern and establishment of a real-time data stream from the station to ETH servers. 2) Processing of received data (streams) using appropriate GNSS software, capable of real-time Precise Point Positioning (PPP). 3) Finally, an initial validation of the acquired tropospheric parameters (e.g. Zenith Total Delay (ZTD)) against post-processing results and GNSS-independent techniques like numerical weather models should be carried out, in order to assess usability for meteorological applications.
Keywords
GNSS, meteorology, zenith total delay, real-time, PPP, low-cost
Labels
Bachelor Thesis , ETH Zurich (ETHZ)
PLEASE LOG IN TO SEE DESCRIPTION
More information
Published since: 2024-05-02 , Earliest start: 2024-09-16 , Latest end: 2024-12-20
Applications limited to ETH Zurich
Organization Space Geodesy (Prof. Soja)
Hosts Soja Benedikt
Topics Engineering and Technology , Earth Sciences
VLBI local baseline observations at the Geodetic Observatory Wettzell
To provide the geodetic infrastructure necessary for monitoring the Earth system and for Global Change research, an end-to-end redesign of the current geodetic VLBI operations called the VLBI Global Observing System (VGOS) is underway. In this framework, new radio telescopes are constructed worldwide. These new telescopes form an independent network that has to be linked to the legacy station network that already has a decade-long observing history. The Geodetic Observatory Wettzell is equipped with three radio telescopes, one 20-meter large legacy antenna that has been operational since 1983 and has contributed the most VLBI measurements worldwide, as well as two new 13-meter large VGOS-style telescopes. By using specially designed VLBI observation sessions between these three telescopes, it is possible to calculate the local baselines between the telescopes with highest accuracy, helping to link the legacy network to the VGOS network.
Keywords
VLBI, scheduling
Labels
Bachelor Thesis , ETH Zurich (ETHZ)
PLEASE LOG IN TO SEE DESCRIPTION
More information
Published since: 2024-05-02 , Earliest start: 2024-09-16 , Latest end: 2024-12-20
Applications limited to ETH Zurich
Organization Space Geodesy (Prof. Soja)
Hosts Soja Benedikt
Topics Engineering and Technology , Earth Sciences
Deriving AuScope VLBI telescope slew models
The AuScope geodetic Very Long Baseline Interferometry (VLBI) array consists of three 12-meter telescopes operated and maintained by the University of Tasmania (UTAS). Two of these telescopes have already joined the operational VLBI Global Observing System (VGOS) and the remaining telescope is expected to join in the next months. The telescopes are of unique design due to the use of a large ball jack screw to change the elevation angle. Therefore, elevation-dependent slewing speeds are expected which are not yet covered in the official antenna slew models that only consist of a rate and a constant overhead time representing acceleration and deceleration.
Keywords
VLBI, telescope slew models, schedule, simulation
Labels
Semester Project , ETH Zurich (ETHZ)
PLEASE LOG IN TO SEE DESCRIPTION
More information
Published since: 2024-05-02 , Earliest start: 2024-09-16 , Latest end: 2024-12-20
Applications limited to ETH Zurich
Organization Space Geodesy (Prof. Soja)
Hosts Soja Benedikt
Topics Engineering and Technology , Earth Sciences
Automated VLBI data analysis pipeline
Over the past four decades, Very Long Baseline Interferometry (VLBI) data analysis has relied heavily on manual processing by various analysts, introducing inherent human biases and leading to inconsistent results. Automating VLBI analysis stands to revolutionize this field by not only simplifying the workflow but also significantly enhancing the quality and consistency of data interpretation.
Keywords
VLBI, data analysis, automation
Labels
Master Thesis , ETH Zurich (ETHZ)
PLEASE LOG IN TO SEE DESCRIPTION
More information
Published since: 2024-05-02 , Earliest start: 2024-08-01 , Latest end: 2025-05-31
Applications limited to ETH Zurich
Organization Space Geodesy (Prof. Soja)
Hosts Soja Benedikt
Topics Engineering and Technology , Earth Sciences
Analyzing and Exploiting Discrepancies in GNSS Solutions of Different Latencies with Machine Learning
The Nevada Geodetic Laboratory (NGL) is a global provider of GNSS solutions, offering data for approximately 20,000 stations worldwide. This includes both daily coordinate estimates and final solutions at a 5-minute resolution. While these final solutions typically exhibit a latency of about 2 weeks, the potential for additional solutions with shorter latency is recognized. Specifically, approximately 9,000 stations are poised for rapid solutions with a latency of around 24 hours, while 640 stations are designated for ultra-rapid solutions with a mere 2-hour latency. This extensive dataset presents a compelling opportunity for the application of machine learning (ML).
Keywords
Machine learning, GNSS, coordinate, troposphere, earth observation data
Labels
Semester Project , ETH Zurich (ETHZ)
PLEASE LOG IN TO SEE DESCRIPTION
More information
Published since: 2024-05-02 , Earliest start: 2024-09-16 , Latest end: 2024-12-20
Applications limited to ETH Zurich
Organization Space Geodesy (Prof. Soja)
Hosts Soja Benedikt
Topics Engineering and Technology , Earth Sciences
Detecting climate change: assessing trends and uncertainties in the GRACE-derived global groundwater product G3P
GRACE/GRACE-FO are the only satellite missions that can globally observe terrestrial water storage (TWS) variations. Extending across two decades, these data provide unique insights into hydrological dynamics and allow for quantification of climate-change related effects on a global scale. Besides the provision of TWS variations, the gravIS (gravity information service) of GFZ Potsdam recently released products of groundwater storage (GWS) variations, which are calculated by subtracting the aggregated and filtered storage contributions for soil moisture, glaciers, snow, and surface water from the GRACE/GRACE-FO-based TWS variations. The GWS dataset was generated within the framework of the G3P project (G3P: Global Gravity-based Groundwater Product, funded by Horizon 2020). Recently, TWS and groundwater have been implemented as Essential Climate Variables by the Global Climate Observing System (GCOS). In the view of climate change detection, this contribution focusses on resolving climate signals at different scales from the GWS time series data. Firstly, a quantitative & qualitative analysis of long-term trends and seasonal signals present in these observations is envisaged. For this, the Hector time series processing software can be used. This analysis includes an assessment of the uncertainty of those data products, using well-defined metrics (e.g., signal-to-noise ratio, power spectral densities). The derived trends should be compared to existing literature, and new groundwater hotspots should be identified. Then, important inter-seasonal variations (such as the El-Nino phenomena), can be extracted using unsupervised learning methods such as principal component analysis, and their contribution to the uncertainty budged can be gauged. Finally, depending on the progress made, the successful candidate can also outline the predictability of these inter-annual variations from meteorological variables, such as precipitation, and their impact on interpretability of groundwater loss (e.g., as observed in the U.S., or India). For this purpose, ML models for the prediction of TWS -- developed at the group of Space Geodesy -- could be tested for applicability.
Keywords
GRACE, GRACE-FO, climate change, terrestrial water storage, machine learning, prediction
Labels
Master Thesis , ETH Zurich (ETHZ)
PLEASE LOG IN TO SEE DESCRIPTION
More information
Published since: 2024-05-02 , Earliest start: 2024-08-01 , Latest end: 2025-05-31
Applications limited to ETH Zurich
Organization Space Geodesy (Prof. Soja)
Hosts Soja Benedikt
Topics Engineering and Technology , Earth Sciences
Calibration of Low-cost GNSS Antennas Using a Robot Arm
The Global Navigation Satellite System (GNSS) serves as a versatile tool in both scientific research and everyday life. However, the Phase Center Offset (PCO) and Variation (PCV) models of GNSS antennas are necessary for high-precision demanding applications, such as accurate attitude determination for nano satellites and the determination of station coordinates at mm-level. The robot arm-based calibration method can achieve absolute antenna calibration with full coverage of azimuth and elevation angles, and thus it is commonly employed for calibration of geodetic GNSS antennas. Nowadays, low-cost GNSS antennas are extensively utilized across various applications owing to their cost-effectiveness. However, such low-cost antennas typically lack the necessary PCO and PCV models, impeding their broader applications. The primary objective of this thesis is to calibrate of several low-cost GNSS antennas using a robot arm located on the rooftop of the HPV building at the ETH Hönggerberg campus. The antennas include products from taoglas and Frauenhofer designed for nano satellites, as well as several smartphone antennas. Additional experiments will be designed to validate and apply the obtained antenna models, with a focus on enhancing attitude determination, positioning accuracy and tropospheric delay determination. Successful completion of this thesis will equip the student the skills to operate the robot arm and a comprehensive understanding of the antenna calibration process.
Keywords
GNSS, antenna calibration, low-cost receiver, robot arm
Labels
Master Thesis , ETH Zurich (ETHZ)
PLEASE LOG IN TO SEE DESCRIPTION
More information
Published since: 2024-05-02 , Earliest start: 2024-08-01 , Latest end: 2025-05-31
Applications limited to ETH Zurich
Organization Space Geodesy (Prof. Soja)
Hosts Soja Benedikt
Topics Engineering and Technology , Earth Sciences
Analysis of Recent Seismic and Tectonic Activity in Greece with the HELLAS GPS Network
Greece is a vital region for Earth monitoring activities due to its complex tectonics and provides a unique opportunity for geodetic and geodynamic studies. For the past three decades, ETH Zurich has actively conducted various scientific activities in Greece, initiated back then by Professor Kahle of the Geodesy and Geodynamics Lab. This includes numerous GPS campaigns and the establishment of approximately 20 GPS permanent stations, with 10 co-located seismometers (HELLAS network). Collaborative initiatives with Greek partners have resulted in several dissertations and numerous research publications. The GPS and seismic data from the stations in Greece are collected in yearly campaigns. However, data recorded by the HELLAS GPS stations since about 2015 have not yet been investigated. This master's thesis focuses thus on the analysis of recent GPS data collected by stations of the HELLAS network in Greece. The primary task involves processing and analyzing the GNSS data from the permanent stations that are still operational. For this purpose, the Bernese GNSS software will be used. First, quality control will be performed to identify potential problems in the GPS data. From the cleaned data, station coordinates and their changes will be estimated, which will reveal valuable insights into tectonic activities and ground motion patterns within the region. Since 2015, several earthquakes with magnitudes greater than 6.0 Mw have occurred. The displacements caused by these earthquakes will be studied in detail based on the GPS measurements. Furthermore, long-term tectonic movements will be investigated. The results will be compared to those from nearby GNSS stations operated by other institutions, including EUREF and IGS stations. This comparison will not only validate the results achieved with the HELLAS data, it will also give an indication of the importance of the HELLAS network for densifying the monitoring networks in that region.
Keywords
GNSS, Greece, seismology, tectonics
Labels
Master Thesis , ETH Zurich (ETHZ)
PLEASE LOG IN TO SEE DESCRIPTION
More information
Published since: 2024-05-02 , Earliest start: 2024-08-01 , Latest end: 2025-05-31
Applications limited to ETH Zurich
Organization Space Geodesy (Prof. Soja)
Hosts Soja Benedikt
Topics Engineering and Technology , Earth Sciences
Detection of Earthquakes from GNSS Displacement Time Series Using Machine Learning
Jumps in GNSS displacement time series (usually named “offsets”) typically result from large earthquakes or equipment changes, and have to be properly introduced in the functional model of these data. When undetected, they can significantly bias the quality of the trends derived, such as tectonic velocities. On the other hand, unreasonably detected offsets (false positives) significantly degrade the uncertainty of estimated trends. Due to the variety of different geophysical and spurious signals present in GPS data (such as hydrological signals, snow cover, or multipath, …), the detection of offsets by well-developed statistical methods still suffers from rather high false positive and/or false negative rates. In turn, recent developments of ML-based detection algorithms (developed at the group of Space Geodesy) show promising results, having precision and recall values of 70-80%. This study should further investigate and improve the performance of ML-driven GNSS offset detection algorithms. The algorithms will be applied to the fully vetted SOPAC GNSS time series data set, provided by Scripps Institution of Oceanography. Besides this, it should be explored which of the above-named spurious signals hinder even higher success rates. Since large earthquakes are rare, it is envisaged to further improve the performance of machine learning by simulating GNSS earthquake data. Finally, the results should be compared against the performance of NASA JPL's TACLS algorithm, provided within the framework of a joint collaboration.
Keywords
GNSS, earthquake detection, machine learning
Labels
Semester Project , ETH Zurich (ETHZ)
PLEASE LOG IN TO SEE DESCRIPTION
More information
Published since: 2024-05-02 , Earliest start: 2024-09-16 , Latest end: 2024-12-20
Applications limited to ETH Zurich
Organization Space Geodesy (Prof. Soja)
Hosts Soja Benedikt
Topics Engineering and Technology , Earth Sciences