With the AI-Lab, we have been creating a space for AI-related teaching and research at the BTU Cottbus-Senftenberg since 2022. In particular for the supervision of theses (B.Sc./M.Sc.) and internships. We employ student assistants for ongoing research projects, which mainly focus on explainable AI in object detection, object tracking, data generation and the solution of differential equations with neural networks. We also offer graduating students with good theses to work together to achieve a publication in form of a conference proceedings or a journal paper.

Equipment

• several high-performance AI workstations
• various lesser powerful computer devices
• a Fischertechnik teaching factory
• two 3D printers
• several wheeled drones (RoboMaster EP/S1)
• several flying drones (RoboMaster TT)
• cameras, microphones and light sources

The AI learning lab is part of the KI@MINT project funded by the Federal Ministry of Education and Research.

Thesis topics (selection/directions)

• Programming the wheeled drones (Bachelor/Master)
• Control of the learning factory based on object detection (Bachlor/Master)
• Detection of process faults in the learning factory (Master)
• Communication wheeled/flying drone (Master)
• Augmentation techniques for training data sets (Bachelor)
• Investigation of object tracking methods (Master)
• Image classification on computers of different power (Bachelor/Master)
• 3D reconstruction with AI (Master)
• Research questions on solving differential equation with AI (Bachelor/Master)
• Reinforcement-Learning in Blender/Unreal (Master)
• Further topics on request

Journal-, Conference and Preprint-Papers

M. Khan Mohammadi, T. Schneidereit, A. Mansouri Yarahmadi, M. Breuß (2024): Investigating training datasets of real and synthetic images for swimmer localisation with YOLO. MDPI AI, 5, pp. 576-593. https://doi.org/10.3390/ai5020030

S. Schneidereit, A. Mansouri Yarahmadi, T. Schneidereit, M. Breuß, M. Gebauer (2024): YOLO-based Object Detection in Industry 4.0 Fischertechnik Model Environment. Intelligent Systems and Applications. Lecture Notes in Networks and Systems, 823, pp. 1-20. doi.org/10.1007/978-3-031-47724-9_1

T. Schneidereit, M. Breuß (2023): Adaptive neural-domain refinement for solving time-dependent differential equations. Advances in Continuous and Discrete Models 2023, 42. doi.org/10.1186/s13662-023-03789-x

Successful theses

Dustin Scharf (study course computer science)
A draft for the implementation of a practical course on Transformer Networks.
Bachelor Thesis, Brandenburg University of Technology, Germany, 2024.

Slavomíra Schneidereit (study course mechanical engineering)
Investigation of object recognition with neural networks using YOLO models in a learning factory.
Master Thesis, Brandenburg University of Technology, Germany, 2022.