Heads of the Research Group: Dr. András OLÁH, Dr. Kálmán TORNAI
Members of the Group: Lóránt DAUBNER, András HALÁSZ, Dániel NÉMETH, Attila TIHANYI, Bálint ÜVEGES
Contact: olah.andras@itk.ppke.hu , tornai.kalman@itk.ppke.hu
Our research focuses on the processing of time-series data originating from sensor-based mobile platforms, wireless sensor networks, and complex information systems. We develop artificial intelligence and machine learning methods for user identification, anomaly detection, intelligent energy applications, and data-driven decision support.
The laboratory was established in 2011. We investigate how measurements collected from mobile devices can be used for behavior-based, soft-biometric user identification on Android and iOS mobile platforms, using both built-in and external sensors. One of our key research directions is open-set recognition, which enables systems not only to identify known users or patterns, but also to recognize unknown, previously unseen behaviors. Beyond mobile biometric identification, the methods developed in this area can also be applied in smart energy systems, smart plug-based appliance recognition, and real-time log data anomaly detection.
The laboratory also conducts research on the industrial and environmental applications of wireless sensor networks. We investigate resilient communication solutions suitable for monitoring environments characterized by hazardous phenomena, such as forest fires, as well as the role of sensor networks in precision agriculture tasks, for example in the monitoring of vineyards.
The goal of the research group is to develop robust AI/ML-based solutions that can be applied in practical environments and support the safer, more efficient, and more intelligent operation of physical and digital systems.
Illustration of how open-set recognition is applied to IMU sensor readings from a smartphone. The proposed method can distinguish among known users and separate their patterns from those of unknown users (i.e., users not seen by the system during training).
Future research directions, collaboration opportunities
The research group offers professional and research collaboration in sensor data pro-cessing, mobile platforms, open-set recognition, and anomaly detection. We work with partners on joint R&D projects, industrial application development, and the preparation of national and international grant proposals.
Key publications
- Halasz, A. P., Al-Hemeary, N., Daubner, L., Juhasz, J., Zsedrovits, .T., Tornai, K. (2025). Adapting a Previously Proposed Open-Set Recognition Method for Time-Series Data: A Biomet-ric User Identification Case Study. ELECTRONICS (SWITZERLAND), 14(20), 3983.
- Németh, D. I., Tornai, K. (2025) . Hybrid ILM–NILM Smart Plug system. INTERNATIONAL JOURNAL OF ELECTRICAL POWER AND ENERGY SYSTEMS, 173 Paper: 111395.
- Üveges, B. Á. and Oláh, A. (2025) . Resilient Multi-Sink, Multipath Routing for Mo-bility-Limited Wireless Sensor Networks in Hostile Event Monitoring. IEEE ACCESS, vol.13, pp.107386-107409.
