CZECH TECHNICAL UNIVERSITY IN PRAGUE
FACULTY OF ELECTRICAL ENGINEERING
DEPARTMENT OF CYBERNETICS
Head of the department:
Prof. Dr. Vladimír Mařík
Karlovo náměstí 13,
121 35 PRAHA 2
About us in PDF.
The Czech Technical University (CTU) in Prague, founded in 1707, is one of the oldest technical universities in the world. It belongs to the leading technical universities in the Czech Republic with approx. 23 000 students enrolled in engineering courses. CTU with over 1500 members of academic staff is also one of the largest research institutions in the Czech Republic. Research is undertaken at CTU in all the basic disciplines taught at the university: i.e. in mathematics, physics, computer science, civil engineering, mechanical engineering, electrical engineering, nuclear and physical engineering, architecture, transportation science, biomedical engineering and in many interdisciplinary areas as well.
The Department of Cybernetics, Faculty of Electrical Engineering (FEE) provides MSc. and postgraduate courses in technical cybernetics, artificial intelligence, computer-integrated manufacturing, computer vision, pattern recognition, and biomedical engineering. The department includes over 65 academic staff and researchers, and over 45 PhD students.
All the activities undertaken in the Department of Cybernetics are directed to achieving the highest standards in research, and to providing a quality, and efficient learning environment at all levels of higher education. In 2000 the Department of Cybernetics, CTU received the award "EU Centre of Excellence" (with the acronym MIRACLE) by the European Commission. The Department became one of the founder members of the newly established Czech Centre of Applied Cybernetics supported through RTD programme of the Czech Ministry of Education.
The research is carried out in two co-operating centres: Gerstner Laboratory for Intelligent Decision Making and Control (GL) and Centre for Machine Perception (CMP), which are recognised as the leading Czech centres for research into computer vision and machine perception, data warehousing, industrial production system integration and production planning information systems, respectively. The centres have gained high international reputation in the respective fields as well.
Head: Vladimír Mařík
The Gerstner Laboratory was founded in 1996 as an extension of the Joint Research Centre of CTU Prague and FAW Linz (Austria). It includes over 40 academic staff and researchers, and over 35 Ph.D. students.
The Gerstner Laboratory carries out basic and applied research in the areas of distributed artificial intelligence, multi-agent systems, machine learning and system diagnostics, advanced database systems and data warehousing, data mining and decision support systems, evolutionary computation, intelligent robotics and biocybernetics.
The laboratory is structured into 4 research groups:
- Agent Technology Group
- Intelligent and Mobile Robotics Group
- Knowledge-Based and Software Systems Group
- Nature Inspired Technology Group
Agent Technology Group
Head: Michal Pěchouček
The research mission of the Agent Technology Group (ATG) is to carry out the leading edge research in the field of theoretical foundations and applications of agent-based computing.
The main strength of the ATG is in prototyping research that is aimed at validation and empirical analysis of the theoretical concepts on computational prototypes of large multi-agent systems. The key concepts of research investigated are social knowledge and acquaintance models, models of trust and reputation in multi-agent systems, modelling and solving agents communication inaccessibility, formal models and detection of agents' adversarial behaviour, distributed decision making, distributed planning, distributed coordination, coalition/alliance formation, negotiation and cooperation, meta-reasoning, monitoring, community intrusion, agents reflection and adjustable autonomy, scalable multi-agent simulations.
In terms of applied research the ATG is mainly active in agent applications in manufacturing (e.g. production planning, control and simulation), supply chain management, logistics but also systems supporting agent-based defense and rescue operations. Recently the ATG also got involved in the domain of collective robotics, where various agent concepts are deployed for the autonomous coordination and deconflicting operations of unmanned land and aerial vehicles.
Intelligent and Mobile Robotics Group
Head: Libor Přeučil
The research conducted by the group is focused mainly on design and development of intelligent and mobile robots. The overall goal is to develop a highly robust cognitive control system able to navigate through and to create and keep a world model at the same time. Reaching this goal leads to solving core sub-tasks: sensing and sensor fusion, mobile robot localization and navigation, world map recovery in 2D and 2.5D, robot activity planning and scheduling for single and multi-robot systems, strategies for collective robot behaviour, etc. Preferably, robust solutions with no extreme demands on hardware implying cheap applications are chosen.
Integration of human and robot entities is also researched. The activities are focused on development of personal assistance and personal navigation systems, Simultaneous Localization and Map Building (SLAM) techniques for human entities, mutual interfacing and knowledge sharing, and communication schemes. Moreover, techniques of cooperation and coordination of multiple systems are studied (e.g. rescue mission planning, cooperative environment mapping, and robot-soccer strategy development).
Other research streams include industrial diagnostics of large systems, safety critical software development for transportation applications and software testing. Research in these topics focuses on support tools for fault-tolerant systems, runtime diagnostics of railway safety and control systems.
Knowledge-Based and Software Systems Group
Head: Zdeněk Kouba
The Knowledge-Based and Software Systems Group is interested in generic aspects of software design, in respect to the design of information systems and knowledge-based systems. The group's main attention is on data warehousing, transforming data between various models, and knowledge management.
In the area of data warehousing the problem of extraction, transformation and load processes (ETL) is being studied. Interesting results have been achieved in the field of interoperability of data warehouses and generic geographical information systems. The generic method of transforming data between different data models is being studied with special focus on transforming data for purposes of populating/updating data warehouses and/or to data pre-processing phase of data different mining processes. The research in knowledge management is aimed at semantic annotation of documents for intelligently managing these documents. Ontology-based collaborative knowledge representation is studied with respect to professional and interest groups.
Nature Inspired Technology Group
Head: Olga Štěpánková
The group's fundamental research is concerned with the design and deployment of adaptive intelligent systems. In particular, the group is developing enhancements to traditional machine learning algorithms for building symbolic models of objects described by attribute values, and studying advanced techniques associated with learning models to explain relations between objects. The group is devising efficient implementations of the learning algorithms by using novel strategies for statistical searching in large spaces of possible models, as well as biologically inspired optimization procedures such as genetic algorithms and neural networks. The group deploys learning and data visualization algorithms as a tool for discovering novel, interpretable knowledge from databases.
The group's applied research is focused on intelligent man-machine interfaces. A patent is pending for I4Control® - a specialized tool allowing someone to control a computer through the movement of the human eye. Dedicated medical decision support systems are being designed and developed in cooperation with medical institutions - most of them combine novel sophisticated methods for processing medical data (e.g. EEG, ECG data) and for knowledge acquisition with efficient knowledge representation and machine learning. Interesting results have also recently been achieved by applying machine learning to discover gene-disease associations from human gene expression data.
Center for Machine Perception
Head: Václav Hlaváč
The Center for Machine Perception (CMP) is a research group active in the fields of computer vision, pattern recognition, and mathematical modelling of uncertainty. The CMP, established in 1996, is a part of the Department of Cybernetics of the Czech Technical University (CTU), Prague. CMP is funded partially by the CTU and a number of national, European and industrial grants. CMP comprises over 25 staff members and over 9 full time Ph.D. students.
CMP main research interests are:
- reconstruction of scenes from multiple images
- omni-directional vision, non-classical cameras
- reconstruction of 3D models from unorganised 3D points
- stereo matching and surface reconstruction
- multi-camera systems for recognition
- object recognition
- medical imaging
- statistical pattern recognition and learning issues
- mathematics of uncertainty, quantum and fuzzy logic
Reconstruction of 3D Scenes
The aim is to reconstruct 3D model of the scene from multiple images. Images from standard projective or wide-angle cameras are used as the data source. Several problems are studied: features for wide-baseline matching, establishing sparse correspondences among images, estimating the position and parameters of cameras, finding dense matching between images, and finally reconstructing the 3D model.
The geometry of wide angle and non-standard omni-directional cameras is studied. Image-based virtual reality representations were developed.
Multi-Camera Systems for Recognition
We developed the Virtual Editor system, which processes multiple video streams and simultaneously analyzes action in the scene and produces a smooth output for a distant observer.
Quantum and Fuzzy Logic
We study different theoretical aspects of fuzzy and quantum logic.
Medical Image Analysis
Topics being studied include: Methods for computer aided diagnosis of thyroid gland diseases, methods for quantification of human breathing movement, measuring temporomandibular joint trajectory, 3D modelling of human face for neuropsychological studies, elastic tissue modelling from ultrasound sequences, deformable object registration from different modalities, brain activity reconstruction from MEG/EEG, lung noduli detection.
Both statistical and structural pattern recognition methods are studied. We study algorithms for large-scale quadratic optimization problems, algorithms for consistent labelling, sequential decision methods for fast object recognition. Results were applied to real-time face detection problem, on-line traffic sign recognition, license plate detection and reading, recognition of activity.