Issued topics of Doctoral Study Program:|
- Analysis and Classification of Textural Features
The thesis is focused on definition and analysis of textural features in images of sedimentary grains acquired by electron microscope device. The main goal of the work is to objectivize and to automatize the estimation of exoscopic parameters of given grains with the purpose to statistically classify the grains into different classes based on a geomorphological processes which contributed to form a given analysed grain structure.
- Analysis of Expressive Music Performance
Musical expression is the art of playing or singing music with emotional communication. Expression include dynamic indications, such as forte or piano, phrasing, differing qualities of timbre and articulation, colour, intensity, energy and excitement. All of these devices are at the service of the composer's intention and they can best be interpreted by the performer. The goal of a dissertation is to analyze the works in terms of a agogics comparing different interpretations, whether they are musicians and conductors. Comparison should include not only different interpretive approaches, but also the historical and social context that it affects.
- Arterial Wall Motion Using Sequential Bayesian Estimation
The thesis is focused on an effective method for an analysis of the movement of the arterial wall through the development of a new methodology for detecting and tracking interest points located at the arterial wall in the ultrasound videosequence. This new methodology will be based on statistical (Bayesian) models of interest point’s evolution and corresponding methods for multi-object tracking. The main aspects of these models and methods are their ability to model and exploit the structural properties of point evolution and to learn these structural properties from the measurement data. (Collaboration partner: Prof. Franz Hlawatsch, TU Wien)
- Communication network quality parameters
The study topic is focused on research methods for measurement and evaluation data parameters of heterogeneous communication networks in terms of end-user access to the Internet and other services.
- Converters for mutual A/D and D/A conversions working in the current mode
The work is focused on the design of A/D and D/A converters working in the current mode. The aim is to design a suitable structure of number-current and current-number converters without internal current-voltage and voltage-current conversions with respect to enhancing the bandwidth in comparison with converters working in the voltage mode. Part of the work is also the design and analysis of current-mode antialiasing filters. The design will start from unconventional circuit element structures such as current conveyors (CCI, CCII, CCIII) with simple or floating output, current feedback amplifiers (CFA) or transconductance amplifiers (OTA, BOTA, DBTA). Requirements to be met by the candidate: the knowledge of circuit theory and simulation programs (MicroCap, PSpice).
- Design of Modern IP Sophisticated Telematic Systems in Transport
Telematic systems are particularly common in transport. Research into telematic systems based on the Internet protocol will be focused on the design of sophisticated, i.e. well-considered, formally well-developed and complicated methods that use IP systems in various areas. Surveillance and protection systems, systems of paying the fare, information systems and interactive applications, etc. are supposed in particular. Localization by GPS, vehicle diagnostics, and vehicle monitoring on ortho-maps in real situations are in the focus. Sophisticated telematic systems will be software simulated, optimized and subsequently hardware realized in the form of functional specimens. Communication between two cars without a driver intervention, collision avoidance, information transmission about traffic from the places cars left are expected. A highly accurate navigation system based on Galileo system (GNSS) for controlling functional blocks of cars are considered.
- Detection and Tracking of Multiple Objects from Multiple Camera Images
The automated detection and tracking of objects based on camera images is an important problem in various surveillance applications. The goal of this PhD thesis is to develop and study methods for the detection and tracking of an unknown number of mobile objects based on multiple camera images. These methods will make use of recently developed techniques of statistical signal processing, including the belief propagation algorithm and multitarget tracking algorithms based on random finite sets. An important aspect of the envisaged work is the modeling of the objects and of their motion. (Collaboration partner: Prof. Franz Hlawatsch, TU Wien)
- Effective Use of IP Networks in Crisis Situations
The aim is to create an effective strategy for the use of the public and private IP network for crisis management. Also propose such a network, which could have the capacity, but also in terms of resistance to ensure the crisis communication. This would, in particular, traffic data, voice, TV broadcast. Other parts would propose new methods of Internet communication management-manage the flows of information, etc. Research should include also the influence of network topology on its stability and security, the speed of the spread of viruses, ability to resist attacks, etc. One of the objectives is to design a software robot that will be able to monitor the network topology, where appropriate, the Internet, the aim is to design a system for exchanging files over the Internet, but without any central control. The system would be intuitively usable. The solution should be safe and allow anonymous of the sender and the recipient of the data. The final. design of highly durable network suitable for crisis situations and this proposal prove theory.
- Feature analysis of current-mode electronic frequency filters
The work is focused on the feature analysis of current-mode electronic frequency filters. The aim is to design algoritmizable methods that lead to finding features of filters generally defined by the schematic. It will be necessary to evaluate the available feature ranges, e.g. quality, pass-band current transfer, dynamic range within a defined supply voltage, sensitivity, etc. In the case of finding the quality range, the point is to find the extremes of a multi-variable non-linear function. Mathematical tools (e.g. Maple or MathCAD) will first be used for this purpose; afterwards the algorithm itself will be designed. Requirements to be met by the candidate: algorithm thinking, the knowledge of computer programming and circuit theory.
- Fractional-Order Electronic Circuits
The topic deals with analog electronic circuits of fractional order. These circuits are characterized by non-integer power of a complex variable in the characteristic equation. The aim is to find new frequency filters, oscillators, synthetic immittance elements, PID controllers and other fractional-order circuits. These circuits will provide more general characteristics compared to their integer-order counterparts. Mathematical description, modeling, implementation and practical use of these circuits are also included.
- More Effective security of networks with IEEE 802.11 Standard
The vulnerability of the IEEE 802.11, attacks on the security methods and the methods for its effectiveness. Wireless network security issues expected standards IEEE 802.11. A detailed introduction to these standards, describe the various types of wireless security, including their properties and the subsequent theoretical analysis. The implementation of the attacks on the various methods of modern security algorithms, emphasis will be placed on the identification of weaknesses of 802.11 protocols. On the basis of the findings will be designed and tested security more effective. Attention will be paid to the security quality of service QoS in 802.11 networks. The network will be subject to examination by appropriate programs and simulators (OMNET++, NS2). Will also be examined according to modern services (IP phones, the transmission of video, voice, multimedia) to delay and to be performed, including an overall analysis to the QoS optimization.
- Parallelization of evolutionary algorithms
The aim is the research in the field of evolutionary algorithms parallelization. Parallelization is an integral part to increase the effectiveness of evolutionary algorithms and their possible use. Research should build on current knowledge and carry out research on the impact of selected parameters and design performance. Research should be concentrated on the island models of evolutionary algorithms. According to the chosen method the knowledge of any programming language is required, as well as scripting language Python or Matlab. Selected development and testing environment is left on free choice. The result should be presented and results verified.
- Perceptual Analysis of Audio Signals
In many areas of audio signal processing, eg. in speech processing, lossy coding of audio signals or in objective assessment of the sound quality, methods based on a model of human hearing are used. The aim of the study is research of these methods for use in other areas, especially for analysis of the sound of musical instruments, for analysis of the listening rooms and for assessment of algorithms for restoration of audio signal, and research of modification of existing models of human hearing for applications in these areas.
- Physical unclonable functions
The study topic is focused on research of physical unclonable functions and their use in cryptographic protocols. Physical unclonable function (PUF) is a physical entity that is embodied in a physical structure and is easy to evaluate but hard to predict. PUF represent an alternative to safely store secret key cryptosystems.
- RC Elements with Distributed Parameters and Their Utilization in Electrical Circuits
The topic aims at the research into design of resistive-capacitive elements with distributed parameters (RC-EDP) and their utilization in electrical circuits especially with modern active elements (conveyors, current amplifiers etc.). The main part of the study will be design of RC-EDP implementing immittance functions of fractional (i.e integer) order. The aim will be improvement of properties compared to present solutions of fractional-order immittance elements (increase in accuracy, frequency bandwidth, better integrability, possibility of parameter setting). Frequency filters, oscillators, synthetic immittance elements and other circuits will be designed as applications of these elements. The aim will be improving properties compared to existing circuits and moreover obtaining characteristics that cannot be realized by classical elements or only at the cost of using complex structures.
- Security in Converged Networks
The aim is to analyse the up-to-date development and trends in the area of converged networks, mainly the problems of protection against cybernetic attacks. The areas of 5G mobile, SDN and consequential transmission technologies are seemed to be among the advanced possibilities. The design of innovative or new protection methods is supposed to be on the basis of obtained observation. Research requires a summary of networks area, practice with MATALB or SCILAB programs, or knowledge of at least one of VHDL, C or Java languages, system FPGA, evolutionary algorithms, etc.
- Sequential Estimation of Arterial Wall Contours
Examination of the common carotid artery (CCA) based on ultrasound measurements is an effective method for detecting cardiovascular diseases. In particular, the spatial evolution of the CCA wall is related to diagnostically relevant parameters describing, e.g., CCA tissue elasticity. Accordingly, the goal of this PhD thesis is to estimate the time-varying contour of the CCA wall from an ultrasound video sequence. This is to be achieved by developing a sequential extension of a machine learning method known as Gaussian process regression (GPR). In the GPR method, a function to be learned is represented by a random process, which avoids an explicit parametric representation. The desired sequential extension of the GPR method should be able to continuously track the detailed shape of the CCA wall over a time period of arbitrary duration. (Collaboration partner: Prof. Franz Hlawatsch, TU Wien)
- Spatiotemporal Analysis of Sound Fields
The spatiotemporal method display the cumulative development of the sound field as a function of direction of the sound intensity using the spatial impulse response. Application of this method is, for example, an analysis of listening rooms, estimation of direction of the incoming sound and more. The aim of this work is to modify this method for various microphone arrays and to find new objective metrics describing the sound field using the spatiotemporal analysis, e.g. spatial information coding and objective assessment of the quality of the listening rooms.
- The design of methods of data detection in xPON systems
The aim is to design methods, an algorithm, to detect data in xPON network. Furthermore, to design the algorithm with a possibility of implementation into FPGA. The focus is on the data detection between OLT and ONU units. An development platform equipped with programmable XILINX network card and G.984 GPON system will be granted. The knowledge of VHDL, FPGA is needed, a knowledge of C or Python is beneficial. The overview of xPON network is an integral part
- Towards Adaptive Optimization of Rehabilitation Processes of Balance Disorders
Brain disorders occur when our brain is damaged or negatively influenced by injury, disease, or health conditions. Balance disorders create a significant category of brain damage impairments. The underlying hypothesis of the considered novel data-centric research approach is that detailed monitoring and continuous analysis of the data produced during patient’s rehabilitation exercise on modern balance force platforms complemented by patient´s physiological data integrated with data captured from other sources can help to optimize the therapy w.r.t. the current needs of the patient, to improve the efficiency of the therapeutic process, and to prevent patient overstressing during the therapy. The aim of this Doctoral Thesis is to develop an adaptive optimization (optimal selection and adjustment of therapy parameters from a large parameter value space) of involved therapeutic processes including precise monitoring, assessment, learning, and response cycles. This research will be conducted in cooperation with researchers from CVUT and Univerzita Karlova Praha (UK), who perform balance training (and collecting appropriate data) in rehabilitation centers built on their two home-grown systems: Homebalance - a system enabling balance training and Scope – a system collecting physiological data. Here, a set of sensors is capturing and storing large sets of data. In the initial phase, data provided by UK will serve as a source of positive and/or negative examples of interventions expected from adaptive optimization.
Thesis advisors (consultants):
• Prof. Peter Brezany, BUT Brno and University of Vienna
• Prof. Olga Stepankova, CVUT Praha
• MUDr. Marketa Janatova, 1. LF Univerzity Karlovy and CVUT Praha
- Transmission of multimedia data in IoT networks
Doctoral thesis deals with transmission of multimedia data in IoT and wireless sensor networks. The aim of a research is to perform a study of possibilities that allowing transmission of multimedia data on devices with limited computing resources and design and development of algorithms ensure transmission of multimedia data with emphasis on high security, low latency and efficient usage of device's resources.
- User identification through Internet traces
All Internet users have certain networking behavior: are visiting some subset of web pages, are members of given groups on social networks, are using another Internet tools like Skype. User obviously keeps his behavior even in case of traveling where the user is presented under various IP addresses. The goal of this work is to study possibilities of user identification based on his/her communication practice. That means e.g. (but not only) visited web pages, duration of staying on given web page, order of web pages visited, etc. It is necessary to take into account varying connection speed and any other parameters which might influence quantitative description of user behavior. Next goal of this work is to estimate the probability of correct user identification based on his/her networking behavior.