Search

Kinematic sensors are often used to analyze movement behaviors in sports and daily activities due to their ease of use and lack of spatial restrictions, unlike video-based motion capturing systems. Still, the generation, and especially the labeling of motion data for specific activities can be time-consuming and costly. Additionally, many models struggle with limited data, which limits their performance in recognizing complex movement patterns. To address those issues, generating synthetic data can help expand the diversity and variability. In this work, we propose IMUDiffusion, a probabilistic diffusion model specifically designed for multivariate time series generation. Our approach enables the generation of high-quality time series sequences which accurately capture the dynamics of human activities. Moreover, by joining our dataset with synthetic data, we achieve a significant improvement in the performance of our baseline human activity classifier. In some cases, we are able to improve the macro F1-score by almost 30%. IMUDiffusion provides a valuable tool for generating realistic human activity movements and enhance the robustness of models in scenarios with limited training data.

In this paper, a thermodynamically consistent solution of the interfacial Riemann problem for the first-order hyperbolic continuum model of Godunov, Peshkov and Romenski (GPR model) is presented. In the presence of phase transition, interfacial physics are governed by molecular interaction on a microscopic scale, beyond the scope of the macroscopic continuum model in the bulk phases. The developed two-phase Riemann solvers tackle this multi-scale problem, by incorporating a local thermodynamic model to predict the interfacial entropy production. Using phenomenological relations of non-equilibrium thermodynamics, interfacial mass and heat fluxes are derived from the entropy production and provide closure at the phase boundary. We employ the proposed Riemann solvers in an efficient sharp interface level-set Ghost-Fluid framework to provide coupling conditions at phase interfaces under phase transition. As a single-phase benchmark, a Rayleigh-B\'enard convection is studied to compare the hyperbolic thermal relaxation formulation of the GPR model against the hyperbolic-parabolic Euler-Fourier system. The novel interfacial Riemann solvers are validated against molecular dynamics simulations of evaporating shock tubes with the Lennard-Jones shifted and truncated potential. On a macroscopic scale, evaporating shock tubes are computed for the material n-Dodecane and compared against Euler-Fourier results. Finally, the efficiency and robustness of the scheme is demonstrated with shock-droplet interaction simulations that involve both phase transfer and surface tension, while featuring severe interface deformations.

We introduce Brain-Artificial Intelligence Interfaces (BAIs) as a new class of Brain-Computer Interfaces (BCIs). Unlike conventional BCIs, which rely on intact cognitive capabilities, BAIs leverage the power of artificial intelligence to replace parts of the neuro-cognitive processing pipeline. BAIs allow users to accomplish complex tasks by providing high-level intentions, while a pre-trained AI agent determines low-level details. This approach enlarges the target audience of BCIs to individuals with cognitive impairments, a population often excluded from the benefits of conventional BCIs. We present the general concept of BAIs and illustrate the potential of this new approach with a Conversational BAI based on EEG. In particular, we show in an experiment with simulated phone conversations that the Conversational BAI enables complex communication without the need to generate language. Our work thus demonstrates, for the first time, the ability of a speech neuroprosthesis to enable fluent communication in realistic scenarios with non-invasive technologies., Comment: 16 pages (39 with supplementary meterial), 6 figures

Adversarial patches threaten visual AI models in the real world. The number of patches in a patch attack is variable and determines the attack's potency in a specific environment. Most existing defenses assume a single patch in the scene, and the multiple patch scenarios are shown to overcome them. This paper presents a model-agnostic defense against patch attacks based on total variation for image resurfacing (TVR). The TVR is an image-cleansing method that processes images to remove probable adversarial regions. TVR can be utilized solely or augmented with a defended model, providing multi-level security for robust prediction. TVR nullifies the influence of patches in a single image scan with no prior assumption on the number of patches in the scene. We validate TVR on the ImageNet-Patch benchmark dataset and with real-world physical objects, demonstrating its ability to mitigate patch attack.

Visual AI systems are vulnerable to natural and synthetic physical corruption in the real-world. Such corruption often arises unexpectedly and alters the model's performance. In recent years, the primary focus has been on adversarial attacks. However, natural corruptions (e.g., snow, fog, dust) are an omnipresent threat to visual AI systems and should be considered equally important. Many existing works propose interesting solutions to train robust models against natural corruption. These works either leverage image augmentations, which come with the additional cost of model training, or place suspicious patches in the scene to design unadversarial examples. In this work, we propose the idea of naturalistic support artifacts (NSA) for robust prediction. The NSAs are shown to be beneficial in scenarios where model parameters are inaccessible and adding artifacts in the scene is feasible. The NSAs are natural looking objects generated through artifact training using DC-GAN to have high visual fidelity in the scene. We test against natural corruptions on the Imagenette dataset and observe the improvement in prediction confidence score by four times. We also demonstrate NSA's capability to increase adversarial accuracy by 8\% on average. Lastly, we qualitatively analyze NSAs using saliency maps to understand how they help improve prediction confidence.

The extraction of text in high quality is essential for text-based document analysis tasks like Document Classification or Named Entity Recognition. Unfortunately, this is not always ensured, as poor scan quality and the resulting artifacts lead to errors in the Optical Character Recognition (OCR) process. Current approaches using Convolutional Neural Networks show promising results for background removal tasks but fail correcting artifacts like pixelation or compression errors. For general images, Transformer backbones are getting integrated more frequently in well-known neural network structures for denoising tasks. In this work, a modified UNet structure using a Swin Transformer backbone is presented to remove typical artifacts in scanned documents. Multi-headed cross-attention skip connections are used to more selectively learn features in respective levels of abstraction. The performance of this approach is examined regarding compression errors, pixelation and random noise. An improvement in text extraction quality with a reduced error rate of up to 53.9% on the synthetic data is archived. The pretrained base-model can be easily adapted to new artifacts. The cross-attention skip connections allow to integrate textual information extracted from the encoder or in form of commands to more selectively control the models outcome. The latter is shown by means of an example application.

The ghost fluid method allows a propagating interface to remain sharp during a numerical simulation. The solution of the Riemann problem at the interface provides proper information to determine interfacial fluxes as well as the velocity of the phase boundary. Then considering two-material problems, the initial states of the Riemann problem belong to different fluids, which may have different equations of states. In the inviscid case, the solution of the multi-fluid Riemann problem is an extension of the classical Riemann problem for a single fluid. The jump of the initial states between different fluids generates waves in both fluids and induces a movement of the interface, similar to a contact discontinuity. More subtle is the extension to viscous and heat conduction terms which is the main focus of this paper. We account for the discontinuous coefficients of viscosity and heat conduction at the multi-fluid interface and derive solutions of the Riemann problem for these parabolic terms, from which we can derive parabolic, interfacial fluxes.

Airfoil icing is a severe safety hazard in aviation and causes power losses on wind turbines. The precise shape of the ice formation is subject to large uncertainties, so uncertainty quantification (UQ) is needed for a reliable prediction of its effects. In this study, we aim to establish a reliable estimate of the effect of icing on airfoil performance through UQ. We use a series of experimentally measured wind tunnel ice shapes as input data. Principal component analysis is employed to construct a set of linearly uncorrelated geometric modes from the data, which serves as random input to the UQ simulation. For uncertainty propagation, non-intrusive polynomial chaos expansion (NIPC), multi-level Monte Carlo (MLMC) and multi-fidelity Monte Carlo control variate (MFMC) methods are employed and compared. As a baseline model, large eddy simulations (LES) are carried out using the discontinuous Galerkin flow solver FLEXI. UQ simulations are carried out with the in-house framework PoUnce. Its focus is on a high level of automation and efficiency considerations in a high performance computing environment. Due to the high number of samples, the simulation tool chain of the baseline model is completely automatized, including a new structured boundary layer grid generator for highly irregular domain shapes. The results show that forces on the airfoil vary considerably due to the uncertain ice shape. All three methods prove to be suited to predict mean and standard deviation. In the Monte Carlo techniques, the choice and performance of low-fidelity models is shown to be decisive for estimator variance reduction. The MFMC method performs best in this study. To our knowledge, there are no UQ studies of iced airfoils based on LES, let alone with advanced UQ methods such as MLMC or MFMC. The present study thus represents a leap in accuracy and level of detail for this application.

The brightest gamma-ray burst ever observed, long-duration GRB 221009A, was detected by GRBAlpha nano-satellite without saturation. We present light curves of the prompt emission in 13 energy bands, from 80 keV to 950 keV, and perform a spectral analysis to calculate the peak flux and peak isotropic-equivalent luminosity. Since the satellite's attitude information is not available for the time of this GRB, more than 200 incident directions were probed in order to find the median luminosity and its systematic uncertainty. We found that the peak flux in the $80-800$ keV range (observer frame) was $F_{\rm{ph}}^{\rm{p}}=1300_{-200}^{+1200}$ ph cm$^{-2}$s$^{-1}$ or $F_{\rm{erg}}^{\rm{p}}=5.7_{-0.7}^{+3.7}\times10^{-4}$ erg cm$^{-2}$s$^{-1}$ and the fluence in the same energy range of the first GRB episode lasting 300 s, which was observable by GRBAlpha, was $S=2.2_{-0.3}^{+1.4}\times10^{-2}$ erg cm$^{-2}$ or $S^{\rm{bol}}=4.9_{-0.5}^{+0.8}\times10^{-2}$ erg cm$^{-2}$ for the extrapolated range of $0.9-8,690$ keV. We infer the isotropic-equivalent released energy of the first GRB episode to be $E_{\rm{iso}}^{\rm{bol}}=2.8_{-0.5}^{+0.8}\times10^{54}$ erg in the $1-10,000$ keV band (rest frame at $z=0.15$). The peak isotropic-equivalent luminosity in the $92-920$ keV range (rest frame) was $L_{\rm{iso}}^{\rm{p}}=3.7_{-0.5}^{+2.5}\times10^{52}$ erg s$^{-1}$ and the bolometric peak isotropic-equivalent luminosity was $L_{\rm{iso}}^{\rm{p,bol}}=8.4_{-1.5}^{+2.5}\times10^{52}$ erg s$^{-1}$ (4 s scale) in the $1-10,000$ keV range (rest frame). The peak emitted energy is $E_p^\ast=E_p(1+z)=1120\pm470$ keV. Our measurement of $L_{\rm{iso}}^{\rm{p,bol}}$ is consistent with the Yonetoku relation. It is possible that, due to the spectral evolution of this GRB and orientation of GRBAlpha at the peak time, the true values of peak flux, fluence, $L_{\rm{iso}}$, and $E_{\rm{iso}}$ are even higher. [abridged], Comment: 7 pages, 7 figures, 1 table, accepted for publication in Astronomy & Astrophysics

(1) Background: The success of physiotherapy depends on the regular and correct performance of movement exercises. A system that automatically evaluates these could support the therapy. Previous approaches in this area rarely rely on Deep Learning methods and do not yet fully use their potential. (2) Methods: Using a measurement system consisting of 17 IMUs, a dataset of four Functional Movement Screening (FMS) exercises is recorded. Exercise execution is evaluated by physiotherapists using the FMS criteria. This dataset is used to train a neural network that assigns the correct FMS score to an exercise repetition. We use an architecture consisting of CNN, LSTM and Dense layers. Based on this framework, we apply various methods to optimize the performance of the network. For the optimization, we perform a extensive hyperparameter optimization. In addition, we are comparing different CNN structures that have been specifically adapted for use with IMU data. Finally, the developed network is trained with the data of different FMS exercises and the performance is compared. (3) Results: The evaluation shows that the presented approach achieves a convincing performance in the classification of unknown repetitions of already known subjects. However, the trained network is yet unable to achieve consistent performance on the data of a previously unknown subjects. Additionally, it can be seen that the performance of the network differs significantly depending on the exercise it is trained for.

The search for research datasets is as important as laborious. Due to the importance of the choice of research data in further research, this decision must be made carefully. Additionally, because of the growing amounts of data in almost all areas, research data is already a central artifact in empirical sciences. Consequentially, research dataset recommendations can beneficially supplement scientific publication searches. We formulated the recommendation task as a retrieval problem by focussing on broad similarities between research datasets and scientific publications. In a multistage approach, initial recommendations were retrieved by the BM25 ranking function and dynamic queries. Subsequently, the initial ranking was re-ranked utilizing click feedback and document embeddings. The proposed system was evaluated live on real user interaction data using the STELLA infrastructure in the LiLAS Lab at CLEF 2021. Our experimental system could efficiently be fine-tuned before the live evaluation by pre-testing the system with a pseudo test collection based on prior user interaction data from the live system. The results indicate that the experimental system outperforms the other participating systems., Comment: Best of 2021 Labs: LiLAS

The Navier-Stokes-Korteweg (NSK) system is a classical diffuse interface model which is based on van der Waals theory of capillarity. Diffuse interface methods have gained much interest to model two-phase flow in porous media. However, for the numerical solution of the NSK equations two major challenges have to be faced. First, an extended numerical stencil is required due to a third-order term in the linear momentum and the total energy equations. In addition, the dispersive contribution in the linear momentum equations prevents the straightforward use of contact angle boundary conditions. Secondly, any real gas equation of state is based on a non-convex Helmholtz free energy potential which may cause the eigenvalues of the Jacobian of the first-order fluxes to become imaginary numbers inside the spinodal region. In this work, a thermodynamically consistent relaxation model is presented which is used to approximate the NSK equations. The model is complimented by thermodynamically consistent non-equilibrium boundary conditions which take contact angle effects into account. Due to the relaxation approach, the contribution of the Korteweg tensor in the linear momentum and total energy equations can be reduced to second-order terms which enables a straightforward implementation of contact angle boundary conditions in a numerical scheme. Moreover, the definition of a modified pressure function enables to formulate first-order fluxes which remain strictly hyperbolic in the entire spinodal region. The present work is a generalization of a previously presented parabolic relaxation model for the isothermal NSK equations.

We present the detector performance and early science results from GRBAlpha, a 1U CubeSat mission, which is a technological pathfinder to a future constellation of nanosatellites monitoring gamma-ray bursts (GRBs). GRBAlpha was launched in March 2021 and operates on a 550 km altitude sun-synchronous orbit. The gamma-ray burst detector onboard GRBAlpha consists of a 75x75x5 mm CsI(Tl) scintillator, read out by a dual-channel multi-pixel photon counter (MPPC) setup. It is sensitive in the ~30-900 keV range. The main goal of GRBAlpha is the in-orbit demonstration of the detector concept, verification of the detector's lifetime, and measurement of the background level on low-Earth orbit, including regions inside the outer Van Allen radiation belt and in the South Atlantic Anomaly. GRBAlpha has already detected five, both long and short, GRBs and two bursts were detected within a time-span of only 8 hours, proving that nanosatellites can be used for routine detection of gamma-ray transients. For one GRB, we were able to obtain a high resolution spectrum and compare it with measurements from the Swift satellite. We find that, due to the variable background, the time fraction of about 67 % of the low-Earth polar orbit is suitable for gamma-ray burst detection. One year after launch, the detector performance is good and the degradation of the MPPC photon counters remains at an acceptable level. The same detector system, but double in size, was launched in January 2022 on VZLUSAT-2 (3U CubeSat). It performs well and already detected three GRBs and two solar flares. Here, we present early results from this mission as well., Comment: 11 pages, 9 figures, submitted to the proceedings book of the conference: SPIE Astronomical Telescopes + Instrumentation 2022

In this paper, we consider Riemann solvers with phase transition effects based on the Euler-Fourier equation system. One exact and two approximate solutions of the two-phase Riemann problem are obtained by modelling the phase transition process via the theory of classical irreversible thermodynamics. Closure is obtained by appropriate Onsager coefficients for evaporation and condensation. We use the proposed Riemann solvers in a sharp-interface level-set ghost fluid method to couple the individual phases with each other. The Riemann solvers are validated against molecular dynamics data of evaporating Lennard-Jones truncated and shifted fluid. We further study the effects of phase transition on a shock-drop interaction with the novel approximate Riemann solvers.

Adversarial attacks in deep learning models, especially for safety-critical systems, are gaining more and more attention in recent years, due to the lack of trust in the security and robustness of AI models. Yet the more primitive adversarial attacks might be physically infeasible or require some resources that are hard to access like the training data, which motivated the emergence of patch attacks. In this survey, we provide a comprehensive overview to cover existing techniques of adversarial patch attacks, aiming to help interested researchers quickly catch up with the progress in this field. We also discuss existing techniques for developing detection and defences against adversarial patches, aiming to help the community better understand this field and its applications in the real world., Comment: A. Sharma and Y. Bian share equal contribution

Synchronisation of wireless inertial measurement units in human movement analysis is often achieved using event-based synchronisation techniques. However, these techniques lack precise event generation and accuracy. An inaccurate synchronisation could lead to large errors in motion estimation and reconstruction and therefore wrong analysis outputs. We propose a novel event-based synchronisation technique based on a magnetic field, which allows sub-sample accuracy. A setup featuring Shimmer3 inertial measurement units is designed to test the approach. The proposed technique shows to be able to synchronise with a maximum offset of below 2.6~ms with sensors measuring at 100~Hz. Also, the results indicate a reliable event generation and detection. The investigated parameters suggest a required synchronisation time of eight seconds. Further research should investigate the temperature changes that the sensors are exposed to during human motion analysis and their influence on the internal time measurement of the sensors. In addition, the approach should be tested using inertial measurement units from different manufacturers to investigate an identified constant offset in the accuracy measurements.

The timing-based localization, which utilize the triangulation principle with the different arrival time of gamma-ray photons, with a fleet of Cubesats is a unique and powerful solution for the future all-sky gamma-ray observation, which is a key for identification of the electromagnetic counterpart of the gravitational wave sources. The Cubesats Applied for MEasuring and Localising Transients (CAMELOT) mission is now being promoted by the Hungarian and Japanese collaboration with a basic concept of the nine Cubesats constellations in low earth orbit. The simulation framework for estimation of the localization capability has been developed including orbital parameters, an algorithm to estimate the expected observed profile of gamma-ray photons, finding the peak of the cross-correlation function, and a statistical method to find a best-fit position and its uncertainty. It is revealed that a degree-scale localization uncertainty can be achieved by the CAMELOT mission concept for bright short gamma-ray bursts, which could be covered by future large field of view ground-based telescopes. The new approach utilizing machine-learning approach is also investigated to make the procedure automated for the future large scale constellations. The trained neural network with 10$^6$ simulated light curves generated by the artificial short burst templates successfully predicts the time-delay of the real light curve and achieves a comparable performance to the cross-correlation algorithm with full automated procedures., Comment: 9 pages, 4 figures, Proceedings of SPIE "Astronomical Telescopes and Instrumentation" 2020

Young nearby runaway stars are suitable to search for their place of origin and possibly associated objects, for example neutron stars. Tetzlaff, Neuh\"auser & Hohle (2011) selected young ($\le 50$ Myr) runaway star candidates from Hipparcos, for which they had estimated the ages from the location in the Hertzsprung-Russell diagram and evolutionary models. Here, we redetermine or constrain their young ages more precisely not only by using the new Gaia DR2 data, but also by measuring lithium, which is a youth indicator. For 308 stars, we took spectra to search for the strong resonance doublet of the lithium-7 isotope at 6708 $\unicode{xC5}$. The spectra were taken with the \'Echelle spectrograph FLECHAS at the University Observatory Jena between February 2015 and June 2018 and with TRES between April 2011 and June 2017 at the Fred L. Whipple Observatory. We found 208 stars with significant occurrence of lithium in their spectra, and five possess a possible age younger or about 50 Myr. Three of these targets are even closer than GJ 182, the nearest known runaway star at about 24 pc. Theses stars are young runaway stars suitable for further investigation of their origin from either a dynamical or supernova ejection., Comment: 34 pages, 30 figures, 9 tables, accepted for publication in AN

Sliding meshes are a powerful method to treat deformed domains in computational fluid dynamics, where different parts of the domain are in relative motion. In this paper, we present an efficient implementation of a sliding mesh method into a discontinuous Galerkin compressible Navier-Stokes solver and its application to a large eddy simulation of a 1-1/2 stage turbine. The method is based on the mortar method and is high-order accurate. It can handle three-dimensional sliding mesh interfaces with various interface shapes. For plane interfaces, which are the most common case, conservativity and free-stream preservation are ensured. We put an emphasis on efficient parallel implementation. Our implementation generates little computational and storage overhead. Inter-node communication via MPI in a dynamically changing mesh topology is reduced to a bare minimum by ensuring a priori information about communication partners and data sorting. We provide performance and scaling results showing the capability of the implementation strategy. Apart from analytical validation computations and convergence results, we present a wall-resolved implicit LES of the 1-1/2 stage Aachen turbine test case as a large scale practical application example.

In this paper we propose a novel regularization strategy for the local discontinuous Galerkin method to solve the Hamilton-Jacobi equation in the context of level-set reinitialization. The novel regularization idea works in analogy to shock-capturing schemes for discontinuous Galerkin methods, which are based on finite volume sub-cells. In this spirit, the local discontinuous Galerkin method is combined with an upwind/downwind finite volume sub-cell discretization, which is applied in areas of low regularity. To ensure the applicability on unstructured meshes, the finite volume discretization is based on a least squares approach., Comment: 31 pages, 6 figures

The Riemann problem is one of the basic building blocks for numerical methods in computational fluid mechanics. Nonetheless, there are still open questions and gaps in theory and modelling for situations with complex thermodynamic behavior. In this series, we compare numerical solutions of the macroscopic flow equations with molecular dynamics simulation data. To enable molecular dynamics for sufficiently large scales in time and space, we selected the truncated and shifted Lennard-Jones potential for which also highly accurate equations of state are available. A comparison of a two-phase Riemann problem is shown, which involves a liquid and a vapor phase, with an undergoing phase transition. The loss of hyperbolicity allows for the occurrence of anomalous wave structures. We successfully compare the molecular dynamics data with two macroscopic numerical solutions obtained by either assuming local phase equilibrium or by imposing a kinetic relation and allowing for metastable states.

Many applications involving complex multi-task problems such as disaster relief, logistics and manufacturing necessitate the deployment and coordination of heterogeneous multi-agent systems due to the sheer number of tasks that must be executed simultaneously. A fundamental requirement for the successful coordination of such systems is leveraging the specialization of each agent within the team. This work presents a Receding Horizon Planning (RHP) framework aimed at scheduling tasks for heterogeneous multi-agent teams in a robust manner. In order to allow for the modular addition and removal of different types of agents to the team, the proposed framework accounts for the capabilities that each agent exhibits (e.g. quadrotors are agile and agnostic to rough terrain but are not suited to transport heavy payloads). An instantiation of the proposed RHP is developed and tested for a search and rescue scenario. Moreover, we present an abstracted search and rescue simulation environment, where a heterogeneous team of agents is deployed to simultaneously explore the environment, find and rescue trapped victims, and extinguish spreading fires as quickly as possible. We validate the effectiveness of our approach through extensive simulations comparing the presented framework with various planning horizons to a greedy task allocation scheme., Comment: Accepted to IFAC World Congress 2020

To improve the understanding of human gait and to facilitate novel developments in gait rehabilitation, the neural correlates of human gait as measured by means of non-invasive electroencephalography (EEG) have been investigated recently. Particularly, gait-related event-related brain potentials (gERPs) may provide information about the functional role of cortical brain regions in human gait control. The purpose of this paper is to explore possible experimental and technical solutions for time-sensitive analysis of human gait-related ERPs during spontaneous and instructed treadmill walking. A solution (HW/SW) for synchronous recording of gait- and EEG data was developed, tested and piloted. The solution consists of a custom-made USB synchronization interface, a time-synchronization module and a data merging module, allowing temporal synchronization of recording devices for time-sensitive extraction of gait markers for analysis of gait-related ERPs and for the training of artificial neural networks. In the present manuscript, the hardware and software components were tested with the following devices: A treadmill with an integrated pressure plate for gait analysis (zebris FDM-T) and an Acticap non-wireless 32-channel EEG-system (Brain Products GmbH). The usability and validity of the developed solution was tested in a pilot study (n = 3 healthy participants, n=3 females, mean age = 22.75 years). Recorded EEG data was segmented and analyzed according to the detected gait markers for the analysis of gait-related ERPs. Finally, EEG periods were used to train a deep learning artificial neural network as classifier of gait phases. The results obtained in this pilot study, although preliminary, support the feasibility of the solution for the application of gait-related EEG analysis..