M.Sc. Carlo Schafflik
M. Sc. Carlo Schafflik
Research Focus
The research focuses on transfer learning approaches for artificial neural networks on embedded systems. The goal is to develop on-chip mechanisms for domain adaptation on low-cost FPGAs in order to enable adaptive AI applications with real-time capability and low energy consumption.
Research of autoencoder-based predictive maintenance for elevators
This work examines the use of autoencoders for predictive maintenance of elevator systems. By analyzing sensor data collected by three acceleration sensors, an autoencoder is used to detect anomalies at an early stage and make maintenance more efficient. The project offers an in-depth exploration of autoencoder architectures and their training strategies for developing accurate anomaly detection. Students have the opportunity to gain practical experience in the field of machine learning and develop their own models. The topic is suitable for bachelor's and master's theses as well as research projects.