Search

Patterns of the magnetic signature of ionospheric currents, generated from an empirical model based on satellite observations, are used to build a statistical correlation based model for ionospheric fields. In order to stabilize the dynamics and to take into account the dominant role of the sun, the fields are represented in solar magnetic coordinates. The covariance structure is analyzed and a second order process that approximates the full dynamics is generated. We show that for synthetic data observations located at the Earth observatories, the full ionospheric field pattern as observed on the earth surface can be reconstructed with good precision. As a proof of principle we provide a first application to the inversion based on real observatory data.

In this work, we present Multiformer, a novel approach to depth-aware video panoptic segmentation (DVPS) based on the mask transformer paradigm. Our method learns object representations that are shared across segmentation, monocular depth estimation, and object tracking subtasks. In contrast to recent unified approaches that progressively refine a common object representation, we propose a hybrid method using task-specific branches within each decoder block, ultimately fusing them into a shared representation at the block interfaces. Extensive experiments on the Cityscapes-DVPS and SemKITTI-DVPS datasets demonstrate that Multiformer achieves state-of-the-art performance across all DVPS metrics, outperforming previous methods by substantial margins. With a ResNet-50 backbone, Multiformer surpasses the previous best result by 3.0 DVPQ points while also improving depth estimation accuracy. Using a Swin-B backbone, Multiformer further improves performance by 4.0 DVPQ points. Multiformer also provides valuable insights into the design of multi-task decoder architectures., Comment: Accepted at the IEEE/CVF Winter Conference on Applications of Computer Vision (WACV) 2025. Code and trained models are available at: https://research.khws.io/multiformer

Pedipulation leverages the feet of legged robots for mobile manipulation, eliminating the need for dedicated robotic arms. While previous works have showcased blind and task-specific pedipulation skills, they fail to account for static and dynamic obstacles in the environment. To address this limitation, we introduce a reinforcement learning-based approach to train a whole-body obstacle-aware policy that tracks foot position commands while simultaneously avoiding obstacles. Despite training the policy in only five different static scenarios in simulation, we show that it generalizes to unknown environments with different numbers and types of obstacles. We analyze the performance of our method through a set of simulation experiments and successfully deploy the learned policy on the ANYmal quadruped, demonstrating its capability to follow foot commands while navigating around static and dynamic obstacles. Videos of the experiments are available at sites.google.com/leggedrobotics.com/perceptive-pedipulation., Comment: Accepted to the IEEE International Conference on Humanoid Robots 2024 Videos available at sites.google.com/leggedrobotics.com/perceptive-pedipulation

Palm oil production has been identified as one of the major drivers of deforestation for tropical countries. To meet supply chain objectives, commodity producers and other stakeholders need timely information of land cover dynamics in their supply shed. However, such data are difficult to obtain from suppliers who may lack digital geographic representations of their supply sheds and production locations. Here we present a "community model," a machine learning model trained on pooled data sourced from many different stakeholders, to produce a map of palm probability at global scale. An advantage of this method is the inclusion of varied inputs, the ability to easily update the model as new training data becomes available and run the model on any year that input imagery is available. Inclusion of diverse data sources into one probability map can help establish a shared understanding across stakeholders on the presence and absence of a land cover or commodity (in this case oil palm). The model predictors are annual composites built from publicly available satellite imagery provided by Sentinel-1, Sentinel-2, and ALOS-2, and terrain data from Jaxa (AW3D30) and Copernicus (GLO-30). We provide map outputs as the probability of palm in a given pixel, to reflect the uncertainty of the underlying state (palm or not palm). This version of this model provides global accuracy estimated to be 92% (at 0.5 probability threshold) on an independent test set. This model, and resulting oil palm probability map products are useful for accurately identifying the geographic footprint of palm cultivation. Used in conjunction with timely deforestation information, this palm model is useful for understanding the risk of continued oil palm plantation expansion in sensitive forest areas., Comment: v03

Germanium is reemerging as a prominent material in the semiconductor field, particularly for electronic applications, photonics, photovoltaics and thermophotovoltaics. Its combination with III-V compound semiconductors through epitaxial growth by metal organic vapor phase epitaxy (MOVPE) is instrumental and thus, the comprehension of the sequential stages in such epitaxial processes is of great importance. During the deposition of GaAs on p-type Ge, the formation of n/p junctions occurs when As diffuses into Ge. It is found that this formation begins in the so-called AsH3 preexposure where Ge substrate is firstly exposed to AsH3. Also important is the fact that both free carrier profiles and As profiles indicate that prolonged AsH3 preexposure times lead to deeper diffusion depths for the same process time. This effect is concomitant with the degradation of the Ge surface morphology, characterized by higher roughness as the AsH3 preexposure duration is extended. Contrary to ion-implanted As in germanium, which shows quadratic dependent diffusivity, our MOVPE investigation using AsH3 indicates a linear relationship, consistent with Takenaka et al.'s MOVPE study using TBAs. Analyzing As profiles alongside simulations, with and without subsequent GaAs epitaxy, suggests the generation of Ge vacancies during the process, contributing to deeper As diffusion.

Classification of road users is important for traffic monitoring. The usability of a height estimate based on the two-ray ground-reflection model as a feature for the classification of vehicles is analyzed in this paper. The four-ray ground-reflection model for fast chirp ramp sequence waveforms of FMCW radars is derived and simplified to the well-known two-ray ground-reflection model. A spectrum from which the height of a target can be derived is obtained using the Lomb-Scargle periodogram. Measurements with two vehicle classes illustrate the approach and show that the model could be used as a feature to distinguish vehicles based on their height., Comment: Preprint of S. Kohnert, D. Zoeke and R. Stolle, "Radar Classification of Vehicles Using a Ground-Reflection Model," 2022 Asia-Pacific Microwave Conference (APMC), Yokohama, Japan, 2022, pp. 641-643, doi: 10.23919/APMC55665.2022.9999980

Target classification is an important task of automotive radar systems. In this work, a concept for estimating the height of vehicles to allow for a differentiation between passenger cars, trucks, and others, is presented and discussed. Fixed installed radar sensors for traffic monitoring in the 77 GHz band are used to track and analyze radar echoes from individual vehicles as they move relative to the radar. Considering multipath propagation, which includes the ground reflection, the height of individual radar targets is estimated by analyzing the periodicity of the resulting amplitude modulation (AM) of the echo signal as a function of the horizontal distance from the radar. Two approaches have been realized to integrate the concept into an automotive FMCW signal processing scheme. Measurements in a test field using a trihedral corner reflector as an idealized target in heights from 0 to 2.5 m suggest, that the concept is suitable for height estimation in an automotive context., Comment: Preprint of S. Kohnert, M. Vogt and R. Stolle, "FMCW Radar Height Estimation of Moving Vehicles by Analyzing Multipath Reflections," 2023 20th European Radar Conference (EuRAD), Berlin, Germany, 2023, pp. 30-33, doi: 10.23919/EuRAD58043.2023.10289187

We used Bands 6 and 7 of the Atacama Large Millimeter/submillimeter Array (ALMA) in Cycles 7 and 8 to search for $\mathrm{CO}\,(3-2)$ emission from a sample of five fast radio burst (FRB) host galaxies discovered by the Commensal Real-time ASKAP Fast Transients (CRAFT) survey and the Fast and Fortunate for FRB Follow-up (F$^4$) team. These galaxies have redshifts $z \approx 0.16-0.48$, masses log$(M_{\rm star}/M_{\odot})\approx 9.30-10.4$ characteristic of field galaxies, and emission lines indicative of ongoing star formation. We detected three of the five galaxies with luminosities $L'(3-2)\approx0.2-4\times10^8\,\rm K\,km \, s^{-1}\,pc^2$ and set upper limits for the other two. Adopting standard metallicity-dependent CO-to-H$_2$ conversion factors, we estimate molecular gas masses $M_{\rm gas}\approx 0.2-3\times 10^9 \, M_{\odot}$. As a population, FRB host galaxies track the main $M_{\rm star}-M_{\rm gas}$ locus of star-forming galaxies in the present-day universe, with gas fractions of $\mu_{\rm gas}\approx0.1$ and gas depletion times $t_{\rm dep} \gtrapprox 1\,$Gyr. We employ the Kaplan-Meier estimator to compare the redshift-corrected $\mu_{\rm gas}$ and $t_{\rm dep}$ for all known FRB hosts with measurements or upper limits with those from the xCOLD GASS survey and find statistically different gas fractions. The difference is not statistically significant when we consider only the five hosts studied here with consistently determined properties, suggesting more FRB hosts with measured molecular gas masses are needed to robustly study the population. Lastly, we present a multi-wavelength analysis of one host (HG20180924B) combining high-spatial resolution imaging and integral field spectroscopy to demonstrate that future high-resolution observations will allow us to study the host galaxy environments local to the FRBs., Comment: 12 pages, 4 figures, 4 tables

Depth-aware video panoptic segmentation is a promising approach to camera based scene understanding. However, the current state-of-the-art methods require costly video annotations and use a complex training pipeline compared to their image-based equivalents. In this paper, we present a new approach titled Unified Perception that achieves state-of-the-art performance without requiring video-based training. Our method employs a simple two-stage cascaded tracking algorithm that (re)uses object embeddings computed in an image-based network. Experimental results on the Cityscapes-DVPS dataset demonstrate that our method achieves an overall DVPQ of 57.1, surpassing state-of-the-art methods. Furthermore, we show that our tracking strategies are effective for long-term object association on KITTI-STEP, achieving an STQ of 59.1 which exceeded the performance of state-of-the-art methods that employ the same backbone network. Code is available at: https://tue-mps.github.io/unipercept

A tandem GaAsP/SiGe solar cell has been developed employing group-IV reverse buffer layers grown on silicon substrates with a subsurface porous layer. Reverse buffer layers facilitate a reduction in the threading dislocation density with limited thicknesses, but ease the appearance of cracks, as observed in previous designs grown on regular Si substrates. In this new design, a porous silicon layer has been incorporated close to the substrate surface. The ductility of this layer helps repress the propagation of cracks, diminishing the problems of low shunt resistance and thus improving solar cell performance. The first results of this new architecture are presented here., Comment: 6 pages, 7 pictures. IEEE PVSC

In today's modern farm, an increasing number of agricultural systems and vehicles are connected to the Internet. While the benefits of networked agricultural machinery are attractive, this technological shift is also creating an environment that is conducive to cyberattacks. While previous research has focused on general cybersecurity concerns in the farming and agricultural industries, minimal research has focused on techniques for identifying security vulnerabilities within actual agricultural systems that could be exploited by cybercriminals. Hence, this paper presents STAVE - a Security Testbed for Agricultural Vehicles and Environments - as a potential solution to assist with the identification of cybersecurity vulnerabilities within commercially available off-the-shelf components used in certain agricultural systems. This paper reports ongoing research efforts to develop and refine the STAVE testbed, along with describing initial cybersecurity experimentation which aims to identify security vulnerabilities within wireless and Controller Area Network (CAN) Bus agricultural vehicle components., Comment: Presented at 17th Midwest Association for Information Systems Conference (MWAIS 2022)At: Omaha, NE, USA

Zinc-diffusion can induce multiple failures in the electrical performance of a multijunction solar cell. In this work, we show an important Zn-diffusion from the AlGaInP back-surface-field layer to the emitter of the GaInP top cell of an inverted multijunction solar cell. Through the analysis of different doping profiles, we provide strong evidence that the diffusion mechanism is (1) triggered by the growth of the tunnel junction cathode and (2) involves point defects. We analyze the implications of Zn-diffusion on the bandgap, the rear-passivation and the minority carrier quality of the GaInP solar subcell by relating the electrical performance of different samples to its corresponding doping profile.

As sensors get more and more integrated, signal processing functions, like tracking, are performed closer to the sensor. Consequently, high level fusion is on the rise. Presented here is a high level fusion scheme incorporating not only linear,but circular quantities as well. Monte Carlo experiments are used to verify our novel fusion operators that work as a weighted average for the Wrapped Normal and the von-Mises distribution. To further verify the new fusion operators, we implemented a full track level fusion scheme and tested it by fusing the measurements of two RADAR sensors.

Infrastructure sensing systems in combination with Infrastructure-to-Vehicle communication can be used to enhance sensor data obtained from the perspective of a vehicle, only. This paper presents a system consisting of a radar sensor network installed at the side of the street, together with an Edge Processing Unit to fuse the data of different sensors. Measurements taken by the demonstrator are shown, the system architecture is discussed, and some lessons learned are presented., Comment: 4 pages, 7 figures