Department of Mathematics, Informatics and Cybernetics

Granting Departments:	Department of Mathematics, Informatics and Cybernetics
Study Programme/Specialization:	Measurement and Signal Processing in Chemistry ( in English language )
Supervisor:	doc. Ing. Pavel Hrnčiřík, Ph.D.

Annotation

In the past decade, there has been a dynamic development of airborne laser scanning of the altimeter of the territory of most European countries, including the Czech Republic. Digital terrain models, which are one of the results of this laser scanning, provide an extremely large amount of detailed information about the nature of the earth's surface within the given territory. Manual analysis of these data sets is very laborious and lengthy and, in the case of examining larger areas, relatively inefficient, especially from the point of view of human labor costs. In this context, machine learning methods offer a promising alternative for solving this very topical problem. This work is specifically focused on the analytical processing of digital terrain models using machine learning methods for the purpose of identifying and classifying the relics of terrain objects created by human activity (use e.g. in archaeology, nature conservation, etc.).

Granting Departments:	Department of Mathematics, Informatics and Cybernetics
Study Programme/Specialization:	Measurement and Signal Processing in Chemistry ( in English language )
Supervisor:	doc. Ing. Dušan Kopecký, Ph.D.

Annotation

The proliferation of modern electronics, integrated circuits, microprocessors and communication and computer technology in general brings with it a high risk of disclosing critical information about the infrastructure in which these elements are used. In the extreme case, there may be a leak or takeover of administrative privileges, which can be misused for digital vandalism, disclosure of important information or attacks on the infrastructure itself. One of the very effective and difficult to detect methods of these attacks is the remote eavesdropping on information that is emanated from electronic devices in the form of electric or magnetic fields. With the development of inexpensive radio technology and as a result of readily available libraries and signal processing algorithms, such an attack may no longer be the sole domain of rich, state-sponsored organizations, but may gradually be adopted by the mainstream hacking community and misused for criminal purposes. The aim of this work is to explore the possibilities and develop and test light and flexible protective shields based on modern nanomaterials, which will serve as an effective passive protection of electronic devices against remote eavesdropping. For this purpose, new composite materials based on electrically conductive nanoparticles with magnetic properties will be prepared. The possibilities of their compatibility with the carrier, chemical structure and morphology, mechanical, electrical and magnetic properties and methods and the possibilities of their processing into the required shape and form suitable for use in miniature electronics will be studied. The experiments will also include testing passive shields in simulated and real conditions and evaluating their ability to dampen electromagnetic waves emitted by electronic devices.

Granting Departments:	Department of Mathematics, Informatics and Cybernetics
Study Programme/Specialization:	Measurement and Signal Processing in Chemistry ( in Czech language )
Supervisor:	prof. Ing. Aleš Procházka, CSc.

Annotation

The dissertation is devoted to motion kinematics based on multi-channel data analysis using computational intelligence and digital multidimensional signal processing tools both in the time, frequency, and scale domains. The methodology includes discrimination methods, machine learning, and pattern vector recognition tools for data classification and motion signals modelling in engineering and biomedicine. The application is devoted to monitoring of physiological signals, recognition of motion patterns, and gait data evaluation using selected wearable sensors including accelerometers, positioning GNSS satellite receivers, and thermal cameras. Results will enable real time data processing for diagnostics and artificial intelligence treatment.

Granting Departments:	Department of Mathematics, Informatics and Cybernetics
Study Programme/Specialization:	Measurement and Signal Processing in Chemistry ( in Czech language )
Supervisor:	doc. Ing. Jaromír Kukal, Ph.D.

Annotation

Granting Departments:	Department of Mathematics, Informatics and Cybernetics
Study Programme/Specialization:	Measurement and Signal Processing in Chemistry ( in English language )
Supervisor:	doc. Ing. Pavel Hrnčiřík, Ph.D.

Annotation

The quality of process control of biotechnological production processes used in the pharmacy and food industry is often constrained by the limited possibilities of on-line measurement of key process parameters (e.g. cell concentration, growth rate, production rate, etc.). One possible solution is the use of software sensors to continually estimate the values of key process indicators from on-line measurable process variables. The proposed PhD thesis is focused on the development of hybrid software sensors and data-driven software sensors with dynamically switched structure, that will be able to evaluate the quality of their estimation during the estimation process and continuously adjust the composition of their data inputs, i.e. use a different on-line measured variable or set of variables for each individual phase of the process.

Granting Departments:	Department of Mathematics, Informatics and Cybernetics
Study Programme/Specialization:	Measurement and Signal Processing in Chemistry ( in Czech language )
Supervisor:	doc. Ing. Jaromír Kukal, Ph.D.

Annotation

Granting Departments:	Department of Mathematics, Informatics and Cybernetics
Study Programme/Specialization:	Measurement and Signal Processing in Chemistry ( in Czech language )
Supervisor:	Ing. Martin Isoz, Ph.D.

Annotation

The work is focused on the application of modern methods of model order reduction (MOR) in engineering practice, including chemical and material engineering. The full order models (FOM) are based on the methods of computational continuum dynamics (both fluids and solids). The FOM-generated data are processed via a posteriori data-driven MOR methods such as proper orthogonal decomposition (POD) or its shifted variant (shiftedPOD). The reduced-order model is prepared either in a standard, projection-based, manner or utilising machine learning. The MOR methodology developed within the dissertation will be applied in a number of engineering-driven optimisation problems.

Granting Departments:	Department of Mathematics, Informatics and Cybernetics
Study Programme/Specialization:	Measurement and Signal Processing in Chemistry ( in English language )
Supervisor:	doc. Ing. Jan Mareš, Ph.D.

Annotation

The aim of the dissertation is the design and implementation of a complex system for the analysis of biomedical data. Data for analysis will be provided/measured at the University Hospital of Královské Vinohrady Prague and the Hospital of the Pardubice Region. The system will (i) serve as an auxiliary tool for the specialist (MD) in the objective assessment of the patient's current condition, (ii) enable the analysis of one- and multi-dimensional data (mainly ECG, heart rate, movement data, possibly CT and NMR). The methodology used for the analysis will be based on classical statistical methods (OLR, RF, etc.) and will also use deep learning methods.

Granting Departments:	Department of Mathematics, Informatics and Cybernetics
Study Programme/Specialization:	Measurement and Signal Processing in Chemistry ( in Czech language )
Supervisor:	prof. Ing. Aleš Procházka, CSc.

Annotation

The dissertation is devoted to multi-channel data acquisition by selected sensor systems and their processing using specific artificial intelligence tools. A special attention is devoted to sensors for simultaneous data acquisition and the use of wireless communication links for their recording and organization in the selected database system. The associated mathematical processing includes data symmetry evaluation, machine learning application, and pattern recognitiion in engineering and biomedicine. The application is devoted to monitoring of neurological signals, evaluation of motion symmetry, and deep learning application for classification of patterns. Results will enable real time data processing using computational intelligence for dynamic access to records through Internet connection.

Granting Departments:	Department of Mathematics, Informatics and Cybernetics
Study Programme/Specialization:	Measurement and Signal Processing in Chemistry ( in English language )
Supervisor:	RNDr. Mgr. Pavel Cejnar, Ph.D.

Annotation

This work concentrates on the processing, reconstruction, and analysis of multidimensional signals, particularly those with significant interfering components. The analysis of mixed chemical samples, utilizing techniques such as mass spectrometry and capillary electrophoresis, generates a vast amount of data, often affected by numerous undesirable physical factors. The objective is to focus on identifying and optimizing suitable statistical and machine learning models. This includes comparing various models and refining them to emphasize the filtering of unwanted components, reconstruction of optimal signals, and direct extraction of significant values. The project involves collaboration with the Department of Biochemistry and Microbiology, leveraging their extensive experience in protein analysis through mass spectrometry.

Granting Departments:	Department of Mathematics, Informatics and Cybernetics
Study Programme/Specialization:	Molecular chemical physics and sensorics ( in English language )
Supervisor:	doc. Ing. Dušan Kopecký, Ph.D.

Annotation

The rapid advent of autonomous systems such as robotic assistants, drones or self-driving vehicles has inevitably brought with it an increase in the use of positioning devices, such as microwave sensors, or advanced lidar, radar or radio technology. This also increases the likelihood of the occurrence of undesired interferences of this electromagnetic wave with the integrated circuits of the autonomous device, which may in turn lead to an increased probability of the occurrence of dangerous phenomena, including accidents and loss of life. The aim of this work is therefore to develop new materials for the attenuation of electromagnetic interference and to apply them as protective shields in the operating area of the electromagnetic spectrum of existing positioning systems. The work will focus on the search, synthesis and characterization of suitable electrical and magnetic materials and their nanostructured analogues and the subsequent design, manufacture and testing of new lightweight and flexible shields. Part of the work will also be modelling and evaluation of the shielding efficiency of protective shields in simulated and real conditions of operation of autonomous systems.

Granting Departments:	Department of Mathematics, Informatics and Cybernetics
Study Programme/Specialization:	Measurement and Signal Processing in Chemistry ( in English language )
Supervisor:	doc. Ing. Dušan Kopecký, Ph.D.

Annotation

Tactile temperature or pressure sensors are devices used in robotics to evaluate the robot's interaction with other objects. These include, for example, manipulating an object, measuring the slip of a gripped object, determining the coordinates of the position of the object or measuring the magnitude of the force acting on the object. The extreme case is complex tactile systems, the purpose of which is to simulate and replace human touch. The sensors used for these purposes must be sufficiently miniature, sensitive to small changes in pressure, must have favorable dynamic properties and time and operational stability of the parameters. Due to the expected high density of tactile sensors connected in simple applications, there must be the possibility of their operation in the form of sensor arrays and data processing using advanced mathematical and statistical algorithms. Last but not least, the cost of producing them must be reasonable so that they can be easily replaced in the event of wear. The aim of this work is therefore to develop new types of tactile temperature and pressure sensors based on modern nanomaterials, which can be used in experiments with the measurement of temporally and spatially distributed forces acting on the matrix of sensors. Part of the work will be the preparation, characterization and processing of thermoelectric and piezoresistive materials based on organic nanostructured semiconductors and carbon nanostructures. Testing of these substances will include, inter alia, structural, chemical and mechanical analysis and measurement of electrical properties in both direct and alternating electric fields. Selected materials will then be processed into sensitive sensors. Part of this work will also be the design of sensor arrays and their testing and signal processing using advanced algorithms.