Endrit Shehaj

Mr. Endrit Shehaj


Research Focus & Metrics

Academic Career Length

10Years (2017-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
11 October 2022 in Acta Astronautica
Verified
Journal Article
Open access
ESEndrit Shehaj
Endrit Shehaj
LMLuca Miotti
Luca Miotti
AGAlain Geiger
Alain Geiger
SDStefano D'ARonco
Stefano D'ARonco
JWJan D. Wegner
Jan D. Wegner
MRMarkus Rothacher
Markus Rothacher

2 December 2022 in Earth System Science Data
Verified
Journal Article
Open access
PDF file
BFBenjamin Fersch
Benjamin Fersch
AWAndreas Wagner
Andreas Wagner
BKBettina Kamm
Bettina Kamm
ESEndrit Shehaj
Endrit Shehaj
ASAndreas Schenk
Andreas Schenk
HKHarald Kunstmann
Harald Kunstmann
Tropospheric water vapor is one of the most important trace gases of the Earth's climate system, and its temporal and spatial distribution is critical for the genesis of clouds and precipitation. Due to the pronounced dynamics of the atmosphere and the nonlinear relation of air temperature and saturated vapor pressure, it is highly variable, which hampers the development of high-resolution and three-dimensional maps of regional extent. With their complementary high temporal and spatial resolutions, Global Navigation Satellite Systems (GNSS) meteorology and Interferometric Synthetic Aperture Radar (InSAR) satellite remote sensing represent a significant alternative to generally sparsely distributed radio sounding observations. In addition, data fusion with collocation and tomographical methods enables the construction of detailed maps in either two or three dimensions. Finally, by assimilation of these observation-derived datasets with dynamical regional atmospheric models, tropospheric water vapor fields can be determined with high spatial and continuous temporal resolution. In the following, a collection of basic and processed datasets, obtained with the above-listed methods, is presented that describes the state and course of atmospheric water vapor for the extent of the GNSS Upper Rhine Graben Network (GURN) region. The dataset contains hourly 2D fields of integrated water vapor (IWV) and 3D fields of water vapor density (WVD) for four multi-week, variable season periods between April 2016 and October 2018 at a spatial resolution of (2.1 km)2. Zenith total delay (ZTD) from GNSS and collocation and refractivities are provided as intermediate products. InSAR (Sentinel-1A/B)-derived double differential slant total delay phases (ddSTDPs) and GNSS-based ZTDs are available for March 2015 to July 2019. The validation of data assimilation with five independent GNSS stations for IWV shows improving Kling–Gupta efficiency (KGE) scores for all seasons, most notably for summer, with collocation data assimilation (KGE = 0.92) versus the open-cycle simulation (KGE = 0.69). The full dataset can be obtained from https://doi.org/10.1594/PANGAEA.936447 (Fersch et al., 2021).

30 November 2023 in Remote Sensing
Verified
Journal Article
Open access
PDF file
JGJunyang Gou
Junyang Gou
CRChristine Rösch
Christine Rösch
ESEndrit Shehaj
Endrit Shehaj
KCKangkang Chen
Kangkang Chen
MSMostafa Kiani Shahvandi
Mostafa Kiani Shahvandi
MRMarkus Rothacher
Markus Rothacher
The International GNSS Service analysis centers provide orbit products of GPS satellites with weekly, daily, and sub-daily latency. The most frequent ultra-rapid products, which include 24 h of orbits derived from observations and 24 h of orbit predictions, are vital for real-time applications. However, the predicted part of the ultra-rapid orbits is less accurate than the estimated part and has deviations of several decimeters with respect to the final products. In this study, we investigate the potential of applying machine-learning (ML) and deep-learning (DL) algorithms to further enhance physics-based orbit predictions. We employed multiple ML/DL algorithms and comprehensively compared the performances of different models. Since the prediction errors of the physics-based propagators accumulate with time and have sequential characteristics, specific sequential modeling algorithms, such as Long Short-Term Memory (LSTM), show superiority. Our approach shows promising results with average improvements of 47% in 3D RMS within the 24-hour prediction interval of the ultra-rapid products. In the end, we applied the orbit predictions improved by LSTM to kinematic precise point positioning and demonstrated the benefits of LSTM-improved orbit predictions for positioning applications. The accuracy of the station coordinates estimated based on these products is improved by 16% on average compared to those using ultra-rapid orbit predictions.

15 August 2017 in Aerospace
Verified
Journal Article
Open access
PDF file
ESEndrit Shehaj
Endrit Shehaj
Vincenzo CapuanoVincenzo Capuano
Vincenzo Capuano
CBCyril Botteron
Cyril Botteron
PBPaul Blunt
Paul Blunt
PFPierre-André Farine
Pierre-André Farine
In recent years, global navigation satellite system (GNSS)-based navigation in high earth orbits (HEOs) has become a field of research interest since it can increase the spacecraft’s autonomy, thereby reducing the operating costs. However, the GNSS availability and the GNSS-based navigation performance for a spacecraft orbiting above the GNSS constellation are strongly constrained by the signals’ power levels at the receiver position and the sensitivity. The simulated level of signal power at the receiver’s position may considerably increase or decrease when assuming different gain/attenuation values of the transmitter antenna for a certain azimuth and elevation. Assuming a slightly different antenna pattern therefore may significantly change the simulated signal’s availability results and accordingly the simulated navigation accuracy, leading to an inexact identification of the requirements for the GNSS receiver. This problem particularly concerns the case of orbital trajectories above the GNSS constellation, where most of the signals received are radiated from the secondary lobe of the transmitters’ antennas, for which typically very little information is known. At the time of this study, it was possible to model quite accurately the global positioning system (GPS) L1 antenna patterns for the IIR and IIR-M Blocks because of the precise information available. No accurate information was available for the GPS L1 antenna patterns of the IIF Block. Even less accurate information was available on the GPS L5 antenna patterns. In this context, this paper aims at investigating the effect of different antenna pattern assumptions on the simulated signal availability and on the consequent simulated navigation performance of a spaceborne receiver orbiting in a very highly elliptical orbit from the Earth to the Moon. Initially the impact of averaging the transmitter’s antenna gain over the azimuth, a typical assumption in many studies, is analyzed. Afterwards, we also consider three different L5 antenna patterns assumed in the literature (the precise L5 patterns are unfortunately not yet fully available). For each of the considered antenna pattern assumptions, we simulate received signal power level, availability, geometric dilution of precision (GDOP), and navigation accuracy in order to evaluate their different effects. After identifying the most conservative assumptions for the transmitters’ antenna patterns, for each elevation of the receiver antenna, we also compute the number of available GNSS observations and analyze their distribution. Moreover, possible aiding of the acquisition process using the prediction of the elevation at which the signal is transmitted, as well as the elevation at which the signal is received, are discussed. Finally, the impact on the GDOP of using only signals transmitted from certain angle intervals of the transmitter antenna pattern and the importance of selecting the transmitters that provide the best GDOP (in the case of a receiver with a limited number of channels) are considered and discussed.

4 November 2020 in NAVIGATION: Journal of the Institute of Navigation
Verified
Journal Article
Open access
PDF file
ESEndrit Shehaj
Endrit Shehaj
KWKarina Wilgan
Karina Wilgan
OFOthmar Frey
Othmar Frey
AGAlain Geiger
Alain Geiger
High spatio‐temporal variability of atmospheric water vapor affects microwave signals of Global Navigation Satellite Systems (GNSS) and Interferometric Synthetic Aperture Radar (InSAR). A better knowledge of the distribution of water vapor improves both GNSS‐ and InSAR‐derived data products. In this work, we present a collocation framework to combine and retrieve zenith and (relative) slant tropospheric delays. GNSS and InSAR meteorological products are combined aiming at a better retrieval of the atmospheric water vapor. We investigate the combination approach with synthetic and real data acquired in the Alpine region of Switzerland. Based on a closed‐loop validation with simulated delays, a few mm accuracy is achieved for the GNSS‐InSAR combination in terms of retrieved ZTDs. Furthermore, when real delays are collocated, the combination results are more congruent with InSAR computed products. This research is a contribution to improve the spatio‐temporal mapping of tropospheric delays by combining GNSS‐derived and InSAR‐derived delays.

7 January 2025 in Atmospheric Measurement Techniques
Verified
Journal Article
Open access
PDF file
ESEndrit Shehaj
Endrit Shehaj
SLStephen Leroy
Stephen Leroy
KCKerri Cahoy
Kerri Cahoy
AGAlain Geiger
Alain Geiger
LCLaura Crocetti
Laura Crocetti
MRMarkus Rothacher
Markus Rothacher
Global Navigation Satellite System (GNSS) radio occultation (RO) is a space-based remote sensing technique that measures the bending angle of GNSS signals as they traverse the Earth's atmosphere. Profiles of the microwave index of refraction can be calculated from the bending angles. High accuracy, long-term stability, and all-weather capability make this technique attractive to meteorologists and climatologists. Meteorologists routinely assimilate RO observations into numerical weather models. RO-based climatologies, however, are complicated to construct as their sampling densities are highly non-uniform and too sparse to resolve synoptic variability in the atmosphere. In this work, we investigate the potential of machine learning (ML) to construct RO climatologies and compare the results of an ML construction with Bayesian interpolation (BI), a state-of-the-art method to generate maps of RO products. We develop a feed-forward neural network applied to Constellation Observing System for Meteorology, Ionosphere, and Climate-2 (COSMIC-2) RO observations and evaluate the performance of BI and ML by analysis of residuals when applied to test data. We also simulate data taken from the atmospheric analyses produced by the European Centre for Medium-Range Weather Forecasts (ECMWF) in order to test the resolving power of BI and ML. Atmospheric temperature, pressure, and water vapor are used to calculate microwave refractivity at 2, 3, 5, 8, 15, and 20 km in geopotential height, with each level representing a different dynamical regime of the atmosphere. The simulated data are the values of microwave refractivity produced by ECMWF at the geolocations of the COSMIC-2 RO constellation, which fall equatorward of 46° in latitude. The maps of refractivity produced using the neural networks better match the true maps produced by ECMWF than maps using BI. The best results are obtained when fusing BI and ML, specifically when applying ML to the post-fit residuals of BI. At the six iso-heights, we obtain post-fit residuals of 10.9, 9.1, 5.3, 1.6, 0.6, and 0.3 N units for BI and 8.7, 6.6, 3.6, 1.1, 0.3, and 0.2 N units for the fused BI&ML. These results are independent of season. The BI&ML method improves the effective horizontal resolution of the posterior longitude–latitude refractivity maps. By projecting the original and the inferred maps at 2 km in iso-height onto spherical harmonics, we find that the BI-only technique can resolve refractivity in the horizontal up to spherical harmonic degree 8, while BI&ML can resolve maps of refractivity using the same input data up to spherical harmonic degree 14.