Goals of this program are to educate most qualified specialists for research and scientific development as well.

Graduates have outstanding knowledge in special area, determined by his/her PhD. project.
Special orientation is given by thesis;graduates are qualified for high positions in all branch.
Leader of research team,managers in industry.

The graduate of the subject field obtains broad knowledge of subject of cybernetics control and/or measuremet. The knowledge is build mainly on theoretical background of the subject. Moreover the graduate will obtain deep special knowledge aimed in direction of his/her thesis. The graduate will be able to perform scientific and/or applied research based on up to date theoretical knowledge. The graduate will be able to organize and lead a team of researchers in the studied subject.

prof. Ing. Pavel Václavek, Ph.D.

2. round (applications submitted from 01.07.2018 to 31.07.2018)

1. Advanced data represenatation, processing and analysis

   Research in the fields of context independent data representation, semantic mapping and process data analysis. The intended research will be focused to utilization of data saved taking advantage of the existing Semantic Sensor Network Ontology for process monitoring purposes, evaluation of temporal constrains of tripplestores with regard to real-time query processing. The research will also focus to data analysis methods aiming to detect non-standard events, process deviations, and other anomalies taking advantage of the stored context in order to reduce false positives.

   Tutor: Fiedler Petr, doc. Ing., Ph.D.

2. Localization of sound and vibration sources with non-contact methods

   The topic is aimed to the research of the methods and algorithms for non-contact localization and characterization of sound and vibration sources. Issues related especially with analysis of the sources with near-field acoustic holography method using microphone array will be studied with respect to its applicability for localization in confined space with reflections and other noise sources and also to increase the prediction accuracy with data fusion from other spatial measurement systems. In addition to the theoretical work, practical implementation of these methods and optimization of calculation algorithms for use in the field of non-contact vibrodiagnostics and localization of noise sources in mechanical systems will be carried out.

   Tutor: Havránek Zdeněk, Ing., Ph.D.

3. Machine Learning Method for Visual Quality Inspection in Industrial Applications

   The topic of the studies is aimed on the usage of machine learning (ML) methods in visual industrial inspection. Specifically, it addresses industrial inspection systems for quality control to the food, beverage, packaging and related industries. Using traditional techniques, like template matching and image comparison, it is difficult to distinguish between image irregularities caused by defects on the surface and image irregularities caused by reflections or by normal surface geometry. ML methods including also deep learning techniques, e.g. convolutional neural networks (CNN) and its several modifications as the Regional CNN, Deep Belief Network and Deep Q-Network, have potential to reliable and precise detection of flaws as scratch, dirt, dent and other defects on the surface of the product without their explicit definition. The main aim of the thesis is to study, design, implement and verify concepts of the ML methods convenient for industrial applications. A necessary requirement for using ML techniques is the availability of a very large set of images for training, which is also part of the thesis solution.

   Tutor: Horák Karel, Ing., Ph.D.

4. Model Order Reduction for Complex Systems Control

   The topic is focused on the problem of model order reduction and related computation complexity reduction of dynamical systems models. The research will deal with methods suitable for linear as well as non-linear systems with respect to preserving system constraints. The goal of the work is to allow application of advanced control methods like MPC for systems, where direct application is not computationally feasible because of high model dimension. The studies will be performed in close relation to international and national research projects in cooperation with industry.

   Tutor: Václavek Pavel, prof. Ing., Ph.D.

5. Modeling and analysis of cyber-physical system with a human operator

   Research focused to elimination of the influence of disturbances introduced into the control system by a human operator due to biophysical characteristics of humans. Research will be focused on measuring and modeling of the movement of the upper limbs in connection with cyber-physical control system by means of electro-mechanical element in order to find a method for separating the active ingredient from the movement disturbances caused by, for example, muscle oscillations. Consequently propose a method to eliminate the disturbances. Part of the research will be focused to quantification of benefits of training, thus assessing the impact of neuro-muscular subsystem on the ability of the operator to implement motion with minimal error component. To verify the model and the proposed methods flight simulators provided by the cooperating workplace (UNOB) will be used.

   Tutor: Bradáč Zdeněk, doc. Ing., Ph.D.

6. Optimal Control of Systems Group

   Design a new method for optimal control of the systems with regard to system throughput, availability and speed of serving and energy consumption. For the design of new control methods perform the research in the field of statistical processing behavior of systems groups and the behavior of people using these groups under varying conditions. Furthermore, perform a research intended to recognize and describe the behavior of people in the system groups. Carry out research into the use of artificial intelligence (fuzzy logic, genetic algorithms, artificial neural networks, etc.) and suggest the optimal method for managing the group management with regard to the defined requirements.

   Tutor: Bradáč Zdeněk, doc. Ing., Ph.D.

7. Resilient software

   Research focused on methods of creating resilient software focusing on embedded systems. Research activities based on methods that are used to design software for applications requiring certification in the field of functional safety (ie. The generation, validation and verification of generated code), assuming that a significant part of R&D activities will focus on usability of available tools for formal specification and code generation for embedded applications including implementation of communication protocols. In the initial phase, it is expected to use Alloy-related tools and functional safety relevant Matlab packages. It is expected that the performed activities will be relevant for ECSEL SECREDAS project.

   Tutor: Fiedler Petr, doc. Ing., Ph.D.

1. round (applications submitted from 01.04.2018 to 15.05.2018)

1. Modelling of dead time LTI SISO systems

   The theme is focused on the modeling of dead time LTI SISO systems using orthogonal functions. As part of his solution the doctoral student will study the properties of generalized Laguerre functions with the parameter generalization. Research will focus on exploiting these properties for the design of dead time dynamic systems models and their implementation and verification of their properties. Solving the topic will take place in relation to national and international projects.

   Tutor: Jura Pavel, prof. Ing., CSc.

2. Modelling of special safety function with respect on industrial Ethernet

   The topic is aimed on research of new special safety functions special safety functions models for machinery and the process safety. The objectives of the thesis consist of the a thorough analysis of the current safety function models, research a thorough analysis of the current models available safety functions, examining the impact of communication, particularly an industrial Ethernet. The student will be designed new models on the base on the analysis and will develop new algorithms for verification of the relevant safety logic functions and security elements for machinery and process safety. The topic will be solved in relation to projects running in cooperation with industrial partners.

   Tutor: Štohl Radek, Ing., Ph.D.

3. Using modern methods for predictive vibrodiagnostics of machines

   Topic is focused on research of advanced methods and algorithms for diagnostics of machines based on their mechanical motion measurement. In particular, the methods used for the advanced identification of measured data and also methods based on the analytical model or previous knowledge will be explored for the identification and classification of fault features. Examples of such methods for pattern classification based on measured data can be auto-associative regression, Mahalanobis-Taguchi strategy or other methods based on artificial intelligence. Research will focus on identifying the practical limitations of these methods and possibilities for their modification, possibly a combination of several methods for achieving greater reliability of state estimation of mechanical system parts (e.g. bearings) or a more successful prediction of failure in the future. In addition to the theoretical work, the practical implementation of these algorithms and their verification on a suitable model is expected. The topic will be solved in conjunction with national and international projects focused on similar research areas.

   Tutor: Havránek Zdeněk, Ing., Ph.D.