Variational Autoencoder to Obtain High Resolution Wind Fields from Reanalysis Data
Rösch, Bernhard; Zacharias, Konstantin; Schlaug, Luca; Westerfeld, Daniel; Geißelsöder, Stefan; Buchele, Alexander (2026)
Rösch, Bernhard; Zacharias, Konstantin; Schlaug, Luca; Westerfeld, Daniel...
WIND (6), 13.
DOI: 10.3390/wind6010013
Open Access Peer Reviewed
Accurate wind flow prediction is essential for various applications, including the placement of wind turbines and a multitude of environmental assessments. Traditionally this can be achieved by using time-consuming computational fluid dynamics (CFD) simulations on reanalysis data. This study explores the performance of an autoencoder (AE) and a variational autoencoder (VAE) in approximating downscaled wind speed and direction using real-world reanalysis data and reference geo- and vegetation data. The AE model was trained for 2000 epochs and demonstrates the ability to replicate wind patterns with a mean absolute error (MAE) of approximately −0.9. However, the AE model exhibited a consistent underestimation of wind speeds and a directional shift of approximately 10 degrees compared to CFD reference simulations. The VAE model produced visually improved results, capturing complex wind flow structures more accurately than the AE model. It mainly achieves better local accuracy and a reduced variance of the results. The overall result suggests that while autoencoders can approximate wind flow patterns, challenges remain in capturing the full variability of wind speeds and directions with sufficient precision. The study highlights the importance of balancing reconstruction accuracy and latent space regularization in VAE models. Future work should focus on optimizing model architecture and training strategies to enhance accuracy, prediction reliability and generalizability across diverse wind conditions and various locations.
Robust Dynamical Component Analysis and Initial Results of Its Application to EEG Data of Epileptic Seizures
Lämmermann, Nina; Warmuth, Monika; Stiehl, Annika; Weeger, Nicolas; Ille, Nicole; Geißelsöder, Stefan; Uhl, Christian (2025)
Lämmermann, Nina; Warmuth, Monika; Stiehl, Annika; Weeger, Nicolas; Ille, Nicole...
47th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC), Copenhagen, Denmark.
DOI: 10.1109/EMBC58623.2025.11253626
Peer Reviewed
Towards Automated EEG-Based Epilepsy Detection Using Deep Convolutional Autoencoders
Stiehl, Annika; Weeger, Nicolas; Uhl, Christian; Bechtold, Dominic; Ille, Nicole; Geißelsöder, Stefan (2025)
Stiehl, Annika; Weeger, Nicolas; Uhl, Christian; Bechtold, Dominic; Ille, Nicole...
47th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC), Copenhagen, Denmark.
DOI: 10.1109/EMBC58623.2025.11254465
Peer Reviewed
Towards Practicable Machine Learning Development Using AI Engineering Blueprints
Weeger, Nicolas; Stiehl, Annika; von Kistowski, Joakim; Geißelsöder, Stefan; Uhl, Christian (2025)
Weeger, Nicolas; Stiehl, Annika; von Kistowski, Joakim; Geißelsöder, Stefan...
IEEE 22nd International Conference on Software Architecture Companion (ICSA-C), Odense, Denmark, 525-528.
DOI: 10.1109/ICSA-C65153.2025.00078
Peer Reviewed
Comparison of Classical EEG Source Analysis with Deep Learning
Winkler, Jakob; Uhl, Christian; Geißelsöder, Stefan; Erdbrügger, Tim; Wolters, Carsten (2025)
Winkler, Jakob; Uhl, Christian; Geißelsöder, Stefan; Erdbrügger, Tim...
In: Aguiar, A.P., Rocha Malonek, P., Pinto, V.H., Fontes, F.A.C.C., Chertovskih, R. (eds) CONTROLO 2024. CONTROLO 2024. Lecture Notes in Electrical Engineering, Springer, Cham 1325, 258–267.
DOI: 10.1007/978-3-031-81724-3_24
Peer Reviewed
This paper discusses the challenges and methods for source reconstruction of evoked potentials using deep learning in the context of electroencephalography (EEG). We propose the use of deep learning to address known challenges and improve traditional approaches. We explain the creation of a suitable dataset for solving the inverse problem, including the simulation of neural activity and the use of lead field matrices for the forward solution. Furthermore, we undertake a comparative analysis of some initial deep learning models with similar classical methods.
Introducing Artificial Neural Networtks to predict the dimensional and micro-geometrial deviations of additively manufactured parts
Vendittoli, Valentina; Polini, Wilma; Walter, Michael S. J.; Geißelsöder, Stefan (2024)
Procedia CIRP 129, 181-186.
DOI: 10.1016/j.procir.2024.10.032
Open Access Peer Reviewed
The dependencies between process parameters and the resulting geometrical accuracy of additively manufactured parts are usually highly non-linear and thus complex to investigate and mathematically quantify. Therefore, the application of artificial intelligence techniques is promising to generate mathematical models that reduce effort and increase the prediction quality. The overall goal is to establish a procedure to automatically determine the optimal settings of the manufacturing process parameters to guarantee the highest geometrical accuracy of parts in additive manufactured production. This paper presents the first step towards this fully automatic procedure – the training and evaluation of a mathematical model based on artificial neural networks to quantify the effects of varying process parameters of a material extrusion process on both macro- and micro-geometrical performances. Therefore, a dataset is established based on the Design of Experiment of an additively manufactured part made from Polylactic Acid filament. The dataset is then used to train an artificial neural network that predicts the dimensional and micro-geometrical deviations of the manufactured parts. Finally, the evaluation of the network's prediction quality and reliability indicate that it is possible to predict the parameters linked to resulting print quality with a mean absolute error from 0.0004 to 0.036.
Using Bayesian Regularized Artificial Neural Networks to Predict the Tensile Strength of Additively Manufactured Polylactic Acid Parts
Vendittoli, Valentina; Polini, Wilma; Walter, Michael S. J.; Geißelsöder, Stefan (2024)
Applied Sciences 14 (8), 3184.
DOI: 10.3390/app14083184
Open Access Peer Reviewed
Additive manufacturing has transformed the production process by enabling the construction of components in a layer-by-layer approach. This study integrates Artificial Neural Networks to explore the nuanced relationship between process parameters and mechanical performance in Fused Filament Fabrication. Using a fractional Taguchi design, seven key process parameters are systematically varied to provide a robust dataset for model training. The resulting model confirms its accuracy in predicting tensile strength. In particular, the mean squared error is 0.002, and the mean absolute error is 0.024. These results significantly advance the understanding of 3D manufactured parts, shedding light on the intricate dynamics between process nuances and mechanical outcomes. Furthermore, they underscore the transformative role of machine learning in precision-driven quality prediction and optimization in additive manufacturing.
Detection of allergens in product information using computer vision
Uhl, M; Knoke, V; Geißelsöder, Stefan (2023)
5th International Conference Business Meets Technology, Valencia, Spain , 241.
Open Access Peer Reviewed
Automatic detection, classification and localization of agricultural operation around wind energy parks
Geißelsöder, Stefan; Madrian, F (2023)
5th International Conference Business Meets Technology, Valencia, Spain.
Open Access Peer Reviewed
Hardware setup for a future autonomous Mobile Manipulation Robot suited for small material handling
Vahlensieck, J; Geißelsöder, Stefan (2023)
5th International Conference Business Meets Technology, Valencia, Spain , 243.
Open Access Peer Reviewed
Kopplung von KI, Strömungssimulation und Strömungsmessung
Zacharias, Konstantin; Welsch, Dennis; Geißelsöder, Stefan; Buchele, Alexander (2023)
mfund Konferenz 2023, Berlin.
Signal analysis and classification of interictal epileptiform discharges from EEG with machine learning
Stiehl, Annika; Geißelsöder, Stefan; Anselstetter, Fabienne; Bornfleth, Harald; Ille, Nicole; Uhl, Christian (2023)
Stiehl, Annika; Geißelsöder, Stefan; Anselstetter, Fabienne; Bornfleth, Harald...
Abstracts of the 57th Annual Meeting of the German Society of Biomedical Engineering 68, 1.
DOI: 10.1515/bmte-2023-2001
Open Access
Topological Analysis of Low Dimensional Phase Space Trajectories of High Dimensional EEG Signals For Classification of Interictal Epileptiform Discharges
Stiehl, Annika; Flammer, M; Anselstetter, Fabienne; Ille, Nicole; Bornfleth, Harald; Geißelsöder, Stefan; Uhl, Christian (2023)
Stiehl, Annika; Flammer, M; Anselstetter, Fabienne; Ille, Nicole; Bornfleth, Harald...
2023 IEEE International Conference on Acoustics, Speech, and Signal Processing Workshops (ICASSPW), Rhodes Island, Greece, 1-5.
DOI: 10.1109/ICASSPW59220.2023.10193167
A new topology based feature extraction method for classification of interictal epileptiform discharges (IEDs) in EEG recordings from patients with epilepsy is proposed. After dimension reduction of the recorded EEG signal, using dynamical component analysis (DyCA) or principal component analysis (PCA), a persistent homology analysis of the resulting phase space trajectories is performed. Features are extracted from the persistent homology analysis and used to train and evaluate a support vector machine (SVM). Classification results based on these persistent features are compared with statistical features of the dimension-reduced signals and combinations of all of these features. Combining the persistent and statistical features improves the results (accuracy 94.7 %) compared to using only statistical feature extraction, whereas applying only persistent features does not achieve sufficient performance. For this classification example the choice of the dimension reduction technique does not significantly influence the classification performance of the algorithm.
Industrielle Anwendungen
Geißelsöder, Stefan (2023)
In: Knowledge Science - Fallstudien: Wie mit Künstlicher Intelligenz die Wissenssicherung und -nutzung im Unternehmen unterstützt wird, Springer Vieweg, Wiesbaden , 193-205.
DOI: 10.1007/978-3-658-41155-8_9
Peer Reviewed
Different montages and dimension reduction methods for EEG signal analysis of Interictal Epileptic Discharges
Stiehl, Annika; Anselstetter, Fabienne; Ille, Nicole; Bornfleth, Harald; Geißelsöder, Stefan; Uhl, Christian (2022)
Stiehl, Annika; Anselstetter, Fabienne; Ille, Nicole; Bornfleth, Harald...
BMT 2022, Abstracts of the 2022 Joint Annual Conference of the Austrian (ÖGBMT), German (VDE DGBMT) and Swiss (SSBE)Societies for Biomedical Engineering 67 (S1), 88.
DOI: 10.1515/bmt-2022-2001
Open Access Peer Reviewed
Dimension reduction methods, persistent homology ans machine learning for EEG signal analysis of Interictal Epileptic Discharges
Stiehl, Annika; Geißelsöder, Stefan; Ille, Nicole; Anselstetter, Fabienne; Bornfleth, Harald; Uhl, Christian (2022)
Stiehl, Annika; Geißelsöder, Stefan; Ille, Nicole; Anselstetter, Fabienne...
Proceedings of the Workshop Biosignale 2022 (24. - 26.08.2022).
DOI: 10.48550/arXiv.2502.12814
Peer Reviewed
On the intelligence of interacting autonomous robots and agents
Geißelsöder, Stefan (2022)
4th International Conference Business Meets Technology, Valencia, Spain.
DOI: 10.4995/BMT2022.2022.15555
Open Access Peer Reviewed
ExDRa: Exploratory Data Science on Federated Raw Data
Baunsgaard, Sebastian; Boehm, Matthias; Chaudhary, Ankit; Derakhshan, Behrouz; Geißelsöder, Stefan; Grulich, Philipp M.; Hildebrand, Michael; Innerebner, Kevin; Markl, Volker; Neubauer, Claus; Osterburg, Sarah; Ovcharenko, Olga; Redyuk, Sergey; Rieger, Tobias; Mahdiraji, Alireza Rezaei; Wrede, Sebastian Benjamin; Zeuch, Steffen (2021)
Baunsgaard, Sebastian; Boehm, Matthias; Chaudhary, Ankit; Derakhshan, Behrouz...
SIGMOD '21: Proceedings of the 2021 International Conference on Management of Data, 2450–2463.
DOI: 10.1145/3448016.345754
Open Access Peer Reviewed
Vorrichtung und Verfahren zur Erkennung von Anomalien in einer industriellen Anlage zur Ausführung eines Produktionsprozesses
Geißelsöder, Stefan; Hitzel, Klaus-Peter; Dilek, Hakan; Hertlein, Christian Klaus; Klose, Marcel Mathias; Tauber, Christian (2020)
Geißelsöder, Stefan; Hitzel, Klaus-Peter; Dilek, Hakan; Hertlein, Christian Klaus...
Europäisches Patent. Internationale Veröffentlichungsnummer: WO2022/069258/A1.
Model-independent search for neutrino sources with the ANTARES neutrino telescope
Albert, A.; André, M.; Anghinolfi, M; Anton, G; Ardid, Miguel; Aubert, J.-J.; et, al; Geißelsöder, Stefan; et, al (2020)
Albert, A.; André, M.; Anghinolfi, M; Anton, G; Ardid, Miguel; Aubert, J.-J.; et, al...
Astroparticle Physics
114, 35-47.
DOI: 10.1016/j.astropartphys.2019.06.003
Open Access Peer Reviewed
Prof. Dr. Stefan Geißelsöder
Hochschule Ansbach
AN[ki]T Zentrum für angewandte KI und Transfer
Fakultät Technik
Residenzstr. 8
91522 Ansbach
T 0981 4877 415 stefan.geisselsoeder[at]hs-ansbach.de
ORCID iD: 0000-0002-9792-0545