Real-time systems research group

04.11.2026.

Introduction

Nowadays, implementing real-time operations in IT applications has become almost unavoidable—not only in extreme cases but also in everyday tasks. What does this mean in practice? A real-time system is one in which the logical correctness of the result depends not only on the outcome of the computation but also on the time at which the result becomes available. The system must perform its tasks within a predetermined time limit (deadline). Such systems include solutions that ensure the continuous movement of characters in a video game, the airbag deployment mechanism during a car crash, or the safety system of a nuclear power plant. Decades ago, in 1987, at the Joint Institute for Nuclear Research in Dubna, we encountered the field of real-time systems, which has remained at the center of our interest ever since and continues to appear in our teaching materials.

In 2017, the research group was founded with the participation of several PhD and MSc students, some of whom have since earned their doctorates and continue to work together on this topic. Students regularly join the ongoing work in this area as well. Current research topics include real-time operating systems, graphical libraries, various forecasting models and methods, and the educational application of devices using such technologies. We introduce MSc students to the exciting world of real-time possibilities through the basic QNX model. Several theses and dissertations have been completed on this topic, and we have also reached out to the public through Researchers’ Night programs. We take great pride in our delta-V demo system—only two universities in Hungary possess such equipment. One of our near-term plans is to involve this device in research, exploring various industrial measurements and control possibilities.

Research Areas

▶ Real-Time Operating Systems

Classification of Operating Systems
Target Systems With or Without an Operating System
Comparison of Preemptive vs. Real-Time Kernels
Real-Time Characteristics
Priorities
CPU Groups
Scheduling

▶ Real-Time Graphics Libraries

We are developing a cross-platform, real-time rendering library for the .NET C# environment, the first working version of which has already been completed. This library supports low-level programming of computer graphics processing units (GPUs) at the abstraction level of graphics APIs, enabling the development of custom rendering algorithms or even complete graphics engines in C#. With our library, shaders, buffers, textures, rendering pipelines, and other well-known low-level graphics resources can be used to implement graphical applications in C#.

➥ System design for the C# graphics API abstraction layer and a project utilizing the library.

The library provides an abstraction layer over low-level graphics APIs, currently supporting Vulkan and several versions of OpenGL. However, each graphics API uses different shader languages. To provide a unified C# abstraction layer for these, the library must also conceal the differences between shader languages. The uniqueness of the development lies in the level of abstraction and its potential applications. The library enables the development of graphics algorithms and graphics engines directly at the level of graphics APIs, in C#. The choice of which graphics API executes the commands is made later—potentially postponed until the moment the program starts. As a result, on every supported platform, the most suitable graphics API available in that environment can always be used.

▶ Forecasting Models

One of the most important tasks in educational environments—such as schools and universities—is predicting student performance, as this contributes to the development of effective mechanisms that, among other benefits, improve academic outcomes and reduce the risk of dropout. During typical student activities, numerous processes operate that handle large amounts of data. These data are often collected and processed automatically by technology-supported learning systems. Through proper analysis and interpretation, valuable insights can be gained into students’ knowledge levels and their attitudes toward learning activities. Such data form the basis for promising algorithms and methods capable of predicting student success. In addition to fulfilling the fundamental objectives of the field, the research introduces several modern methodologies widely applicable to performance prediction, such as machine learning techniques. The aim of the research is to implement and compare existing approaches to develop an optimized solution and to create an accurate model capable of reliably forecasting student performance.

➥ Student’s Demographic and Geographic Feature Identification (SDGFI).

▶ Real-Time Educational Applications and Smart Devices

Students’ expectations regarding education have changed significantly: they value features familiar from social media, knowledge sharing, interactive collaboration with instructors, and instant answers to their questions—even during lectures. However, the applications available on the market have not proven flexible enough for our needs, so we developed our own two-way lecture management system. The real-time operation of the web application is ensured by the SignalR module. The students responded very positively to the new solution, and our measurements confirmed that the learning process became more effective. During the COVID lockdown, we also introduced another real-time application, whose efficiency and alignment with student expectations were likewise examined.

Methodology

The world of real-time systems is heterogeneous. It does not have general support comparable to today’s operating systems (Unix-based or Windows systems). Although there are Unix-based systems adapted to this field, they do not provide a universal solution to the problem. The focus of the research is on how different problem domains can ensure real-time functionality through their specific solutions.

Research Group Members

Dr. Zoltán Illés, Professor (research group leader, MTMT, Tud-O-Méter)
Dr. Viktória Bakonyi, Associate Professor
Dr. Chaman Verma, Assistant Professor
Dr. Dávid Szabó, Assistant Professor
Dr. Szilárd Korom
and many students

Grants and Projects

2024–2027: FITPEDGAI, Future IT Professional Education [54 000 EUR]
2017–2021: EFOP-3.6.3, Valós idejű rendszerek vizsgálata [52M HUF]

Current Active Partnerships

Amity University, university network (joint research, student exchange)
UCW (University Canada West) (joint research, emerging student exchange)
Evosoft (educational collaboration)

Dissertations

Bakonyi Viktória: Valós-idejű oktatási alkalmazások okoseszközök használatával
Verma Chaman: Identification of student’s demographic, geographic features and opinions towards ICT and mobile technology
Szabó Dávid: Valós idejű grafika multi-platform .NET C# környezetben
Korom Szilárd: A rendszerszintű gondolkodás informatikai kompetencia fejlesztése

Selected Publications

C. Verma, Z. Illés, V. Stoffová, V. Bakonyi (2020): Opinion Prediction of Hungarian Students for Real-Time E-Learning Systems: A Futuristic Sustainable Technology-Based Solution, Sustainability [DOI] Q1
C. Verma, Z. Illés, V. Stoffová, V. H. Bakonyi (2021): Comparative Study of Technology With Student's Perceptions in Indian and Hungarian Universities for Real-Time: Preliminary Results, IEEE Access [DOI] Q1
C. Verma, Z. Illés, D. Kumar (2024): TCLPI: Machine Learning-Driven Framework for Hybrid Learning Mode Identification, IEEE Access [DOI] Q1
V. H. Bakonyi, Z. Illés (2018): Real Time Classroom Systems in Teachers Training, Lecture Notes in Computer Science [DOI] Q2
C. Verma, Z. Illés, D. Kumar (2024): An Investigation of Novel Features for Predicting Student Happiness in Hybrid Learning Platforms – An Exploration Using Experiments on Trace Data, International Journal of Information Management Data Insights [DOI] D1

Contact

Dr. Zoltán Illés – illes@inf.elte.hu