List of proposed projects
available as 2025-2026 semester projects
Supervisor: Dr David Harvey / Ethan Tregidga
Type of Project: Semester project (can be extended for a Master thesis)
Duration: 14 weeks (Official start/end date: September 8-December 19)
Submission of final report: January 8
Final Presentation: TBD
Recommended: This project is suitable for a student interested in machine learning, cosmology, dark matter, gravitational lensing.
Context: Dark matter makes up most of the matter in the universe, yet its nature remains one of the biggest open questions in physics. Machine learning now presents and exciting way to probe data in a new and efficient way and answer some of the biggest questions.
Project: In this project, you’ll build a self-supervised machine learning “foundation model” that learns from both cosmological simulations and real telescope data, with the goal of probing dark matter through galaxy clusters.
By capturing general features that work across multiple datasets, the model will be reusable for many other astronomy and cosmology problems, from detecting rare objects to classifying large-scale structures.
The project will involve deep learning, contrastive learning, and multimodal data, applying these methods to large datasets.
This research will contribute to the development of scalable flexible tools for next-generation astronomical surveys, enabling deeper insights into the nature of dark matter and large scale structure.
References
- https://academic.oup.com/mnras/article/531/4/4990/7697182
- https://academic.oup.com/mnras/article/527/3/7459/7452889
- https://arxiv.org/abs/2503.15312
- https://iopscience.iop.org/article/10.3847/2041-8213/abf2c7
- https://openreview.net/forum?id=w0iQslghXD
Supervisor: Dr David Harvey / Ethan Tregidga
Type of Project: Semester project (can be extended for a Master thesis)
Duration: 14 weeks (Official start/end date: September 8-December 19)
Submission of final report: January 8
Final Presentation: TBD
Recommended: This project is suitable for a student interested in machine learning, cosmology, dark matter, gravitational lensing.
Context: Dark matter makes up most of the matter in the universe, yet its nature remains one of the biggest open questions in physics. Machine learning now presents and exciting way to probe data in a new and efficient way and answer some of the biggest questions.
Project: In this project, you’ll build a self-supervised machine learning “foundation model” that learns from both cosmological simulations and real telescope data, with the goal of probing dark matter through galaxy clusters.
By capturing general features that work across multiple datasets, the model will be reusable for many other astronomy and cosmology problems, from detecting rare objects to classifying large-scale structures.
The project will involve deep learning, contrastive learning, and multimodal data, applying these methods to large datasets.
This research will contribute to the development of scalable flexible tools for next-generation astronomical surveys, enabling deeper insights into the nature of dark matter and large scale structure.
References
- https://arxiv.org/abs/2210.02747
- https://arxiv.org/abs/2507.11842
- https://arxiv.org/abs/2307.08698
Supervisor: eSpace/Laboratory TBD (Prof. Emma Tolley/Chris Finlay/Stephan Hellmich) with external support from CU Boulder (Prof. Kevin Gifford)
Type of Project: Semester project (can be extended for a Master thesis)
Duration: 14 weeks (Official start/end date: September 8-December 19)
Submission of final report: January 8
Final Presentation: TBD
Recommended: This project is suitable for a student interested in radio astronomy/interferometry, space situational awareness, space policy, radio interferometry data reduction, and API development/web scraping.
CONTEXT
The steeply increasing number of satellites in orbit is causing more and more interference with astronomical observations. Images acquired by optical telescopes are increasingly affected by satellite trails that can make parts of the observation unusable. For radio observations, the situation is even worse. Strong RFI (Radio Frequency Interference) from satellite communication can ruin a whole observation while weak interference, for example from low-power radio frequency signals that is leaking from the onboard electronics is difficult to detect and needs to be removed from the data in a very computationally expensive process. In the past, contaminated radio astronomy data was flagged and removed before further processing. However, the large and increasing number of satellites in orbit make this approach more costly with respect to telescope time. It is still not the standard for astronomers to schedule their observations to avoid bright satellites crossing the field of view. While for optical telescopes mainly orbital data is required to dodge satellites, for radio telescopes also spectral data (the frequency bands used by the satellites for communication) is very important. Orbital data for most satellites is made publicly available by satellite tracking networks such as space-track, LEOLABS or by the satellite operators themselves. Spectral data is usually not published. Operators are registering for certain frequency bands with the International Telecommunication Union (ITU) but not all of these allocations are publicly available. In addition to this, frequently detected out of band transmissions show that operators do not always comply with their allocated frequencies. Further, there is no information on the spectral profiles of the transmissions, which is very important to efficiently remove RFI from radio astronomy data.
PROJECT SCOPE
The goal of this project is to support radio astronomy with valuable information for observation planning and RFI removal. Unlike for orbital data, there currently is no complete and publicly available database for spectral data. Radio astronomers started to collect such data from various sources on the internet in order to build up such a repository. This project aims to support this activity by incorporating data from the ITU databases that contain the frequency band allocations for all satellites. Further, there exist several publicly available astronomical archives that contain RFI. Identifying and analyzing contaminated data in these archives might allow astronomers to better constrain the spectral profile of the transmissions and provide valuable information for RFI removal.
TASKS
- Identify which data can be shared by the ITU.
- Develop tools to incorporate ITU data into existing RF repositories.
- Perform a survey on publicly available radio astronomy data.
- Develop a tool that can identify datasets that are likely to contain useful data for extracting satellite spectral profiles.
- If time permits, develop a radio interferometry data processing tool to extract frequency spectra for individual satellites.
Supervision: eSpace/LASTRO (Prof. Jean-Paul Kneib/Stephan Hellmich)
with external support from TU Delft (Dominic Dirkx)
Type of Project: Master thesis project
Duration: 18/26 weeks (Official start/end date: TBD)
Submission of final report: TBD
Final Presentation: TBD
Recommended: This project is suitable for a student interested in orbital mechanics, atmospheric modelling and space sustainability.
CONTEXT
During the last five years, the number of satellites in orbit has dramatically increased due to the satellite mega constellations that are currently installed in low Earth orbit (LEO). The high pace in the development of new satellite communication technologies results in mega constellation satellites being frequently replaced which in turn leads to the number of objects that reenter Earth’s atmosphere significantly increases. To reduce ground casualty risks, the satellites are designed for optimal demise which results in almost the entire mass of the satellites being dispersed in the atmosphere. The substances released during the demise have consequences that need to be quantified in order to understand the implications of the increasing number of reentries. This project aims to improve the accuracy of reentry trajectories in order to enable dedicated observations of satellite demise that are required to quantify the implications on the atmosphere.
PROJECT SCOPE
Most reentries are uncontrolled which means that the exact location is not known and makes dedicated observations of the breakup and demise impossible. During the last few orbits, the trajectory of a LEO satellite is increasingly influenced by atmospheric drag. Precise propagation of the trajectory thus relies on information of the shape and attitude of the satellite as well as timely atmospheric data on density and wind speeds in the atmosphere. This project aims to incorporate this information in orbit propagation in order to increase the accuracy of the predicted reentry location. It is planned to implement these capabilities as into the open source astrodynamics library “TU Delft Astrodynamics Toolbox” Tudat , developed at TUDelft. Tudat already contains basic functionality to consider object shape and attitude as well as information on the atmosphere using the NRLMSISE-00 global reference atmospheric model. The main objective of this project is to implement a more accurate atmospheric model that can incorporate real-time weather data and provides more comprehensive data required to determine the precise reentry point.
Tasks
- Familiarize yourself with Tudat
- Perform a literature review of available atmospheric models that can be implemented
- Identify and implement the most suitable model
- Setup an orbit propagation example with shape and attitude dependent drag and lift coefficients
- Use data from historic reentry events to evaluate the new method
Literature
Proposed by: Andrii Neronov, Ettore Zaffaroni, Ronald Scaria
Cosmic rays entering Earth atmosphere produce Extensive Air Showers (EAS) of high-energy particles that can can be observed by an array of particle detectors on the ground. This phenomenon has been discovered back in 1938 by Pierre Auger at Jungrfaujoch laboratory in Switzerland. The EAS detection technique is nowadays used to study the highest energy particles known in nature (ultra-high-energy cosmic rays reaching 100 Exa-electronvolt) and for observations of the highest energy gamma-rays (highest energy gamma-rays reach Peta-electronvolt). One possible type of simple particle detectors is a water tank equipped with a fast photosensor, a Photo-Multiplier Tube (PMT). High-energy particles passing through the tank with the speed faster than speed-of-light in water emit blue Cherenkov light that can be detected with PMT.
The goal of the work is to participate in development of a prototype Water Cherenkov Detector Array (WCDA) in Geneva Lake for observations of EAS with energies in the Tera-electronvolt to Peta-Electronvolt. This prototype will be used for a feasibility study of a large WCDA for measurement of cosmic ray electron spectrum up to highest energies and improvement of sensitivity of observations of gamma-ray sky in the Tera-electron to Peta-electronvolt energy range.
Proposed by: Andrii Neronov
Magnetic fields that are relic of epochs right after the Big Bang can still reside in the low density regions of the Large Scale Structure, between galaxies and galaxy clusters. Detection of these fields and measurement of their properties might provide a valuable “window” on physical processes that have operated in the Early Universe a fraction of a second after the Big Bang. Such measurement is possible with the methods of gamma-ray and radio astronomies, using new observational facilities: gamma-ray Cherenkov Telescope Array (CTA) and radio Square Kilometer Array (SKA). The gamma-ray measurement technique is based on the observation of extended glow around distant extragalactic sources produced by electromagnetic cascade developing along the gamma-ray beam during its propagation through the intergalactic medium. Radio technique is based on the observation of Faraday rotation of the polarised radio signal through the intergalactic medium.
The goal of the work is to get an overview of possible mechanisms of generation of magnetic fields during the first microsecond after the Big Bang, their evolution toward the present-day intergalactic magnetic field and to study a possibility to constrain the “cosmological magnetogenesis” scenario with CTA and SKA observations.
Proposed by: Andrii Neronov, Volodymyr Savchenko
The highest energy gamma-rays produced by sources like supernova remnants, pulsar wind nebulae, active galactic nuclei, interact in the Earth atmosphere and produce extensive air showers of high-energy particles Such showers appear as very fast “shooting stars” on the sky, if imaged at Giga-Hertz imaging rate with imaging atmospheric Cherenkov telescopes. A new system of telescopes based on this technique, Cherenkov Telescope Array Observatory (CTAO) is now in construciton and first Large Size Telescopes (LST) of CTAO are starting to take data.
The goal of the project is to get aquainted with the principle of operation of CTAO, methods of analysis of data, to analyze the first data of LST, for a selection of sources and to get an idea of the mechanisms involved in acceleration of high-energy particles and production of gamma-rays in these sources.
Proposed by: Andrii Neronov, Volodymyr Savchenko
Cherenkov Telescope Array Observatory is a major next-generation gamma-ray telescope that is in construction now. The first Large Size Telescope (LST) of CTAO is already operational and takes data, detecting gamma-rays through the Cherenkov glow of high-energy particle showers initiated by gamma-rays penetrating the Earth atmosphere.
The project is aimed at understanding the methods of gamma-ray data taking and data analysis with CTAO, with a focus on the study of violently variable gamma-ray emission from “blazars”, a special type of Active Galactic Nuclei with supermassive black holes ejecting high-energy particle jets aligned along the line of sight.
This project can be done as a TP-IVb, other 8 ECTS, or Master project.
Contact: Andrii Neronov, Volodymyr Savchenko
Proposed by: Aris Tritsis (Postdoc)
Magnetic fields are recognised to play a central role in the formation of stars and planets. While their exact influence on the evolution of molecular clouds — the birthplaces of stars — remains an active area of research, it is widely accepted that magnetic fields reduce the star formation efficiency by at least a factor of a few by supporting these clouds against their self-gravity. Observations show that dense filamentary structures within molecular clouds tend to align perpendicular to the magnetic field, whereas lower-density, linear, semi-periodic structures in the outskirts of these clouds (known as striations) are consistently aligned with the magnetic field. The formation of striations has been attributed to compressible magnetohydrodynamic (MHD) waves.
The goal of this Master’s project is to investigate the correlation between the wavelength of the waves that produce striations and the spacing between prestellar and protostellar cores within dense filaments. This will provide insights into the processes responsible for molecular cloud fragmentation and star formation. The student will use archival data from the Herschel Space Observatory and carry out numerical MHD simulations. By the end of the project, the student will gain hands-on experience with Python/Fortran programming and statistical data analysis.
Proposed by: Mark Sargent
Galaxy emission in the radio domain (GHz frequencies/cm wavelengths) contains information about the current and past growth of key galaxy components (stellar populations, the central supermassive black hole), as well as the reservoir of interstellar gas that will facilitate the future growth of these galaxy building blocks. Long assumed to be characterised by simple power laws, recent studies have revealed that the shape of GHz radio spectra in fact display significant variability between different galaxies (some having, respectively, steep/flat/curved spectra). The physical causes for this are not well understood.
The aim of this project is to (i) build a data base of multi-frequency radio observations for the ~300 galaxies in the JINGLE [1] galaxy sample, and then (ii) utilize this data base to establish which galaxy properties correlate most directly with shape variations of their GHz radio spectra. The galaxies in the JINGLE sample are esp. well suited to this study as the physical properties expected to impact radio spectra (e.g., star-formation rate, stellar mass, size, chemical composition, and dust and gas content) have been accurately measured. The project outcomes will provide important lessons for planning observations with future radio telescopes, and pave the way to a more robust physical charactersation of the galaxies detected in current/future radio surveys.
Skills you will acquire during this project:
– astrophysics knowledge: galaxy formation and evolution, radiation/emission processes
– data processing and visualization
– data analysis (model fitting, resampling methods, statistical treatment of missing data…)
– programming and workflow development
– accessing, cross-referencing and managing data bases
– radio astronomy methods and observation planning
… and more!
Type of project: Master project/Semester project
Contact: Mark Sargent, admin LASTRO
Proposed by: Mark Sargent
The intensity of the radio synchrotron continuum emission from star-forming regions in galaxies can be used to measure the rate at which new stars are being born. As the complex astrophysics involved makes calibration from first principles difficult, these radio synchrotron-based star-formation rate (SFR) measurements are empirically calibrated against other SFR estimators. Due the benefits of the radio SFR measurements technique (e.g., no bias from dust absorption) a lot of effort is being invested into improving our calibration of the technique for different types of galaxies and cosmic epochs.
The goal of this project is to develop reliable approaches for calibrating radio SFR measurements for particularly radio-faint galaxies (e.g., nearby low-mass objects or very distant galaxies). For faint galaxy populations calibration is often attempted in a sample-averaged sense, rather than on a galaxy-by-galaxy basis. As part of this project you will explore which biases may arise in this case in this noise-dominated regime – and how to identify the factors that cause such biases – in order to be able to derive prescriptions for how to correct them, and consequently improve the accuracy of radio SFR measurements. You will use a combination of mock and real data for this. The outcomes of this project will inform survey strategies for future radio observations with, e.g., the SKA [1].
Skills you will acquire during this project:
– astrophysics knowledge: galaxy formation and evolution, radiation/emission processes
– data processing and visualization
– programming and workflow development
– data analysis (statistical treatment of non-detections, resampling methods, …)
– machine learning algorithms
– radio astronomy methods and observation planning
… and more!
Type of project: Master project/Semester project
Contact: Mark Sargent, admin LASTRO
Proposed by: Ashutosh Mishra (PhD) & Emma Tolley
Cosmology with Cold DM has been extensively studied in simulations, but observational predictions for alternate models of DM have not been explored as extensively. This is especially true for bosonic or Fruzzy Dark Matter (FDM), where nonlinear interference effects can leave a distinct signature in the matter power spectrum and small scale structure. Unfortunately, the numerical resolution requirements to faithfully follow the FDM dynamics are much harder to fulfil than for the N-body techniques applicable in the CDM case. A new class of neural networks use physics-based constraints to solve the data limitations and generalization problems of traditional neural networks, called Physics-Informed Neural Networks (PINNs). Such neural networks can be constrained to respect any symmetry, invariance, or conservation principles originating from the physical laws that govern the observed data, as modeled by general time-dependent and non-linear partial differential equations. Embedding this prior knowledge allows scientists to develop higher fidelity networks which require much less training data.
FDM obeys the Shroedinger-Poisson equations, which can be embedded in a PINN framework to create fast simulations of cosmological FDM. We are interested in exploring applications of this technique with Quantum neural networks following the framework developed in [1].
[1] https://arxiv.org/pdf/2209.14754
Proposed by: David Harvey (Faculty), Yves Revaz (Faculty)
Simulating the formation of galaxies is vital to our understanding how the Universe formed and the underlying physics behind it. By simulating galaxies we can test theories, and probe the nature of the elusive dark matter. However, simulating galaxies can take a long time, in particular, owing to the treatement of the cooling due to molecules and metals. Indeed, an accurate treatement requires to solve many non-linear inter-dependent equations. If we can speed up the estimation of metal-cooling in galaxies we will be able to dramatically speed up our simulations. In a complete new and unique way, this masters project will aim to use deep learning to quickly and precisely estimate the abundances and cooling of a set of atomic species at play during the formation of galaxies. By constructing a model that can bypass the need to carry out complicated equations we can hopefully speed up the simulations and hence open up the possibility to probe new models of dark matter. The student will get then hands on simulations, machine learning, deep learning and gain experience in using python packages such as tensor flow and the GRACKLE libraries.