In the Model Making Lab, students can join application-oriented research projects in which the theoretical research problems are inspired by direct practical applications. This often leads to complex research of an interdisciplinary nature, which requires not only IT and mathematical knowledge, but also application-specific knowledge and cooperation with specialists in different fields. That is why several university lecturers participate in the work of the lab, who coordinate and help the students' work in various research areas, such as:

• signal and image processing
• machine learning, artificial intelligence
• geometric modeling, computer graphics
• numerical methods

In the lab, it is possible to deepen the knowledge acquired during the BSc/MSc studies, as well as to prepare TDK and thesis/diploma work, as well as to obtain a professional internship.

The Model Making Lab (IPM-22modKVMTNM1L, IPM-22AUTAUSLAB1) is a 5-credit course that, in addition to model making, other specialties, e.g. students of the Master's program in autonomous Systems Informatics can also enroll.

Technologies used: MatLab, Python, PyTorch, C++, NVidia Falcor, HLSL

Instructors: Dr. Péter Kovács, Dr. Gergő Bognár, Dr. Gábor Fábián, Tamás Dózsa, Csaba Bálint

Contact person: Dr. Péter Kovács, kovika@inf.elte.hu

Website: https://modelinglab.inf.elte.hu

PROJECTS

• Medical Signal Processing:
  • Segmentation of ECG signals, classification of heartbeats, data compression.
  • Processing EEG signals, epilepsy detection, sleep phase classification.
  • Investigation of multi-source human biological signals: estimating blood pressure from ECG and PPG signals, analysis of various medical databases.
• Technical Signal Processing:
  • Autonomous vehicle control: classification of tire sensor signals, road type classification, estimation of tire contact forces, detection of slippage, modeling of power steering control.
  • Telecommunication: simulation of OFDM communication, simulation of transmitter and receiver, channel modeling, model-based decoding, physical layer encryption.
  • System identification: study of dynamic systems and approximation of their transfer functions, research on rational function approximation methods, hardware-oriented implementation algorithms, e.g., on microcontrollers, FPGAs.
• Tomographic Methods:
  • Thermographic image reconstruction: non-destructive material testing.
  • CT image reconstruction: image enhancement, product detection, segmentation.
• Model-Driven Machine Learning
  • Integration of mathematical models with various machine learning algorithms.
  • Development of adaptive projection networks (VPNet), ODE Network, WaveletKernelNet, OptNet, application of Wiener-Hammerstein type networks.
• Geometric Modeling
  • Collision detection of polygons, partitioning of object models, computer representations of 3D surfaces, analysis and development of special data structures.
  • Generation of Voronoi diagrams for polygons, modeling of distance functions.
• Computer Graphics
  • Study of ray tracing algorithms, rendering of implicit surfaces.
  • Shading and shadow calculations.
  • Use of NVidia Falcor for efficient real-time rendering.

COOPERATIONS

• Johannes Kepler University Linz
• Kepler University Hospital
• Silicon Austria Labs
• RECENDT Research Center for Non-Destructive Testing GMBH
• Josef Ressel Center, University of Applied Sciences, Upper Austria
• Faculty of Engineering, RITEH University of Rijeka
• Nanosensors Laboratory, Centre for Energy Research, ELKH - MFA
• Systems and Control Laboratory, Institute for Computer Science and Control, ELKH - SZTAKI
• Department of Computational Sciences, Wigner Research Centre for Physics