275003 research outputs found
Sort by
Open-Source Platform for Integrating High-Reporting rate Information Using FIWARE Technology
No full textThe rapid growth of the Internet of Things (IoT) has resulted in an exponential increase in the availability of real-time data from various sources. However, integrating and processing of high-reporting rate information poses significant challenges due to the large volume, communication, and heterogeneity of the data. In this paper, we are proposing an innovative framework providing a scalable and flexible architecture that enables efficient handling of high-reporting rate data streams from diverse equipment towards pre-processing, storage, visualization and third-party applications integration using the FIWARE technology
Pushing the Boundaries of Dynamic Flavor Analysis with PTR-MS
No full text3350Proton transfer reaction-mass spectrometry (PTR-MS) is an analytical technique that detects volatile organic compounds (VOCs), amongst them flavor/aroma compounds, in the gas-phase. The instrument is constructed and configured in a manner that allows rapid and continuous compound detection over a broad dynamic range. These features have positioned PTR-MS as an ideal measurement tool in food science to characterize dynamic processes, such as flavor release or food spoilage. Since it emerged on the analytical chemistry landscape more than a quarter century ago, the capabilities of PTR-MS have been exploited in numerous food/flavor research applications to complement the conventional approach of gas chromatography-mass spectrometry (GC-MS). PTR-MS boasts a comprehensive back catalog of studies on all manner of foods and beverages, ranging from coffee, to fruits, meat, dairy and oils. Assessments span volatile (flavor) composition screening, dynamic flavor release, spoilage, fermentation or maturation process monitoring, high-throughput profiling, model food systems, and in vivo release, amongst others. This chapter presents an overview of the development and uses of PTR-MS in food/flavor research, from its historic first steps to present day advanced applications.140
ON-BOARD IMAGE PROCESSING WITH FPGA ACCELERATION USING DEEP NEURAL NETWORK INFERENCE
No full textThe acquisition and processing of earth observation data from satellites is in transition. New Earth observation constellations of small satellites generate large amounts of data that cannot easily be transmitted to the ground. In the near future, the amount of data is expected to significantly grow, as sensor resolution and realizable constellation sizes are increasing much faster than the downlink capacity. The processing power of small satellites which rely on commercial off-the-shelf (COTS) hardware is nowadays capable of processing large amounts of data efficiently on board, despite the limited resources available. Consequently, this paper proposes to exploit the power of current advances in deep-learning-based image processing to extract relevant data from the acquired images, which is then transferred to the ground. This contribution attempts to demonstrate the feasibility of using state-of-the-art machine-learning image analysis, directly aboard a small satellite. Thus, on-board pre-processing can make better use of the available communication capacity by discarding unimportant data and reducing the latency between acquisition and reception of information. This results in a reduction of operating costs and increases satellite autonomy. To demonstrate this concept, a deep convolutional neural network has been trained and deployed to a Zynq UltraScale+ MPSoC-based data processing unit (DPU) that is being developed at the Fraunhofer Ernst-Mach-Institut (EMI) for the use in nanosatellite imaging payloads. This DPU contains a Field Programmable Gate Array (FPGA) that allows the implementation of a hardware accelerator, enabling an efficient and fast inference of neural networks. The convolutional neural network employed is a semantic segmentation model trained on satellite imagery. This model uses a UNET-type architecture to segment satellite images into 12 surface classes (e.g., water, vegetation, clouds). It is trained utilizing publicly available ground truth data from the Sentinel-2 L2A mission. To evaluate the model's performance, it is afterwards tested with data from different satellites with comparable spectra and ground resolution. The proposed approach demonstrates that the employed DPU can apply a complex neural network on the typical data volume of a small earth observation satellite in real-time. Furthermore, it demonstrates that this performance can be achieved despite the limited power and cooling resources available on board a small satellite. The use of deep neural network for onboard pre-processing enables the imminent extraction of information from earth observation data and can help the satellite operator to react to short-term events and to determine which data is worth transmitting to the ground
Findings of the WMT 2023 Biomedical Translation Shared Task: Evaluation of ChatGPT 3.5 as a Comparison System
No full text4354We present an overview of the Biomedical Translation Task that was part of the Eighth Conference on Machine Translation (WMT23). The aim of the task was the automatic translation of biomedical abstracts from the PubMed database. It included twelve language directions, namely, French, Spanish, Portuguese, Italian, German, and Russian, from and into English. We received submissions from 18 systems and for all the test sets that we released. Our comparison system was based on ChatGPT 3.5 and performed very well in comparison to many of the submissions
Thermoplastic Polyurethanes: Versatile Shape Memory Polymers for Innovative Applications
No full textShape memory polymers are a class of intelligent materials that have become increasingly important in recent years. Particularly noteworthy are thermoplastic polyurethanes, which are among the best-researched representatives of this group. These materials are characterized by their exceptional versatility, as they can be synthesized with a wide range of properties. This allows them to be adapted to specific application requirements, making them valuable in numerous industrial and technical fields. A notable advantage of thermoplastic polyurethanes is the targeted control over the thermally activated, one-way shape memory effect. This is achieved through both selected chemical formulations and a thermomechanical treatment known as programming. Programming enables the material to return to its original shape after thermal exposure. In four-dimensional (4D) printing, an extension of conventional three-dimensional printing, the temporal dimension is integrated in addition to three-dimensional shaping through the introduction of internal stresses. The advantages of this technology are manifold. 4D printing enables the production of objects that can change their shape or properties in response to environmental conditions.
The use of shape memory polymers also enables more efficient designs that increase functionality. In addition, 4D printing promotes the development of intelligent and interactive products that can respond flexibly to the needs of users. In summary, the presentation provides insights into the synthesis, processing and 4D printing of polyurethanes with shape memory properties and highlights promising technological developments that open up new design possibilities from an application perspective
Efficient in-situ reliability monitoring of in-plane NED Bending Actuators during accelerated aging
No full text525528Fraunhofer IPMS develops since 2015 [1] NED bending actuators, which are integrated in many different applications, such as micro-positioning platforms, micro machined ultrasonic transducers, micro speakers and so on. All the applications need to be analyzed from the reliability point of view. For that reason different tests are performed to force the accelerated aging of the structures, e.g. tests in the climatic chamber. During the accelerated aging the actuator could for example change their resonance frequency or deflection could drop. Big and quick deflections, as by ultrasonic applications result also in a high damping, thus a certain design adjustments may need to be implemented. To analyze all these effects the frequency dependent acquisition of the actuator movements need to be performed. But it is often impossible, because of the measure setup. During the accelerated aging the chips lays typically in a climatic chamber or during the damping tests in a vacuum chamber, so no in-situ measurements are possible. In such cases the analysis of the electrical impedance response can bring the necessary information. For that reason, a simple analytical impedance response model has been developed and verified, which helps to monitor efficiently the changes in the actuator behavior during the stress tests
Filament Development for Metallic Fused Filament Fabrication of Aluminium Alloys
No full textMaterial Extrusion with filaments, Fused Filament Fabrication - FFF, is the most widespread additive manufacturing technology. This counts mainly for polymers, since the use of this technology for metal and ceramic powders is still in its early days. For several metals, filaments can be bought in the market. Aluminium has a low sintering temperature already in the range of the temperatures for thermal debinding of many feedstocks and needs special sintering regimes. To meet these special requirements, a feedstock for the production of highly filled aluminium filaments was developed. The printed parts are debinded by solvents. Here, the following results are presented: 1) rheological behaviour of the feedstock, 2) debinding behaviour with emphasis on the selection of solvents and 3) printing performance
UCRAID (Ukrainian Citizen and refugee electronic support in Respiratory diseases, Allergy, Immunology and Dermatology) action plan
No full text25812595Eight million Ukrainians have taken refuge in the European Union. Many have asthma and/or allergic rhinitis and/or urticaria, and around 100,000 may have a severe disease. Cultural and language barriers are a major obstacle to appropriate management. Two widely available mHealth apps, MASK-air® (Mobile Airways Sentinel NetworK) for the management of rhinitis and asthma and CRUSE® (Chronic Urticaria Self Evaluation) for patients with chronic spontaneous urticaria, were updated to include Ukrainian versions that make the documented information available to treating physicians in their own language. The Ukrainian patients fill in the questionnaires and daily symptom-medication scores for asthma, rhinitis (MASK-air) or urticaria (CRUSE) in Ukrainian. Then, following the GDPR, patients grant their physician access to the app by scanning a QR code displayed on the physician's computer enabling the physician to read the app contents in his/her own language. This service is available freely. It takes less than a minute to show patient data to the physician in the physician's web browser. UCRAID—developed by ARIA (Allergic Rhinitis and its Impact on Asthma) and UCARE (Urticaria Centers of Reference and Excellence)—is under the auspices of the Ukraine Ministry of Health as well as European (European Academy of Allergy and Clinical immunology, EAACI, European Respiratory Society, ERS, European Society of Dermatologic Research, ESDR) and national societies.781
Guided Stochastic Optimization for Motion Planning
No full textLearning from Demonstration (LfD) is a family of methods used to teach robots specific tasks. It is used to assist them with the increasing difficulty of performing manipulation tasks in a scalable manner. The state-of-the-art in collaborative robots allows for simple LfD approaches that can handle limited parameter changes of a task. These methods however typically approach the problem from a control perspective and therefore are tied to specific robot platforms. In contrast, this paper proposes a novel motion planning approach that combines the benefits of LfD approaches with generic motion planning that can provide robustness to the planning process as well as scaling task learning both in number of tasks and number of robot platforms. Specifically, it introduces Dynamical Movement Primitives (DMPs) based LfD as initial trajectories for the Stochastic Optimization for Motion Planning (STOMP) framework. This allows for successful task execution even when the task parameters and the environment change. Moreover, the proposed approach allows for skill transfer between robots. In this case a task is demonstrated to one robot via kinesthetic teaching and can be successfully executed by a different robot. The proposed approach, coined Guided Stochastic Optimization for Motion Planning (GSTOMP) is evaluated extensively using two different manipulator systems in simulation and in real conditions. Results show that GSTOMP improves task success compared to simple LfD approaches employed by the state-of-the-art collaborative robots. Moreover, it is shown that transferring skills is feasible and with good performance. Finally, the proposed approach is compared against a plethora of state-of-the-art motion planners. The results show that the motion planning performance is comparable or better than the state-of-the-art.
Real-World Audio Deepfake Detection Using SSL-Based Speech Models and Diverse Training Data
No full text14651469The potential for audio deepfakes to be used for malevolent purposes is increasing in line with advances in artificial intelligence and synthesis methods. With this, the need for reliable audio deepfake detectors increased. Most audio deepfake detectors comprise two principal components: a frontend, which is responsible for feature extraction, and a back-end, which performs the classification. Self-supervised learning (SSL) based front-ends are right now the most promising when faced with real-world data. We tested different combinations of six SSL-based front-ends and four back-ends, i.e. classifiers, using nine variously combined training sets, enabling the inclusion of the majority of the currently available training sets for audio deepfake detection. The combination of Wav2Vec2.0 XLS-R (2b) as the front-end and GF as the back-end performed best with an EER of 0.73% on the in-the-wild dataset, outperforming the current SOTA. Furthermore, our findings highlighted the significance of training set constellations and the utilisation of large front-ends