|
|
Neuronové sítě a evoluční algoritmy
Vágnerová, Jitka ; Rychtárik, Milan (oponent) ; Hrubeš, Jan (vedoucí práce)
Diplomová práce se zabývá použitím zvolených evolučních algoritmů k určení a úpravě parametrů neuronové sítě. K úpravě parametrů sítě se zpětným šířením chyby byly použity genetické algoritmy, evoluční strategie a evoluční programování. Součástí práce je program vytvořený v prostředí Matlab, ve kterém byly použité metody testovány na úlohách rozpoznávání vzorů a predikci průběhu funkce. Výsledkem práce jsou grafy průběhu chyby sítě a fitness během úpravy pomocí zvolených algoritmů a průběhů chyby při následném učení.
|
|
|
Parallel Computing and Neural Networks in Behavioral Modeling
Vágnerová, Jitka ; Škvor,, Zbyněk (oponent) ; Dědková, Jarmila (oponent) ; Lukeš, Zbyněk (vedoucí práce)
This thesis is focused on methods for the aircraft equipment modeling. The first part provides a brief overview of classical system modeling approaches used for system description, identification, and modeling. Then adaptive, fuzzy and hybrid methods used mainly for black-box system modeling are introduced. Aim of the thesis is to develop an algorithm for identification and modeling of a general system, which can be nonlinear, dynamic and complex. Multiple inputs and multiple outputs of model are assumed. The main part of the thesis introduces a new method which falls into the hybrid systems. It combines fuzzy approach with parametrically defined rules and general regression neural network. Firstly, the fundamentals of simple general regression neural network and its smoothness parameter determination are presented. Secondly, the general regression neural network with the fuzzy rules is introduced. Third part of the thesis is focused on the parallel computing, one of the main objectives. The final algorithm is designed for the parallel machine, because the computational time can be significantly high and for the larger datasets, the model is not achievable when evaluated in single thread. Block diagram for parallel computing in Matlab and CUDA is provided, as well as the basic structure of CUDA source code. Finally, the method is verified on data obtained from the measurement of a miniaturized aircraft model using the antenna outside the aircraft and the probe inside the fuselage of the aircraft model. The validation of the method is done using mean squared error and compared to mean squared error of corresponding model performed using the multilayer neural network with backpropagation learning and Levenberg-Marquardt algorithm.
|
|
|
LHC Abort Gap Monitor acquisition system
Pacner, Petr ; Vágnerová, Jitka (oponent) ; Kubíček, Michal (vedoucí práce)
The Beam Dump System of the LHC Large Hadron Collider is one of its critical systems. It is responsible for changing the trajectory of the beam towards the beam dump. For its functionality it uses an Abort Gap - a space with no particles within a beam structure. This gap is formed during the injection of the particles into the LHC. The injection is never ideal hence Abort Gap always contains some particles. Energy accumulated in the Abort Gap can cause various damages and so its amount has to be continuously estimated. This is the purpose of an Abort Gap Monitor. The currently deployed Abort Gap Monitor uses analog integrators to measure the particle population. This makes the instrument difficult to calibrate due to its hardware inaccuracies. To improve the system measurement versatility and reproducibility the effort is made to create an acquisition system with digital integration. This work aims to design, verify and implement the digital acquisition system. The digital acquisition system serves as one of the major parts of the new CERN SY-BI group Abort Gap Monitor. This document is split into four sections. The first briefly describes the structure of the Beam Dump System and the importance of the Abort Gap. The second section explains the principle of Abort Gap measurements and the data evaluation methodology. The third section describes the current system design and its issues. Then it focuses on the new Abort Gap Monitor topology followed by acquisition system development. The last section presents the simulations and measurements obtained with the new Abort Gap Monitor Acquisition System and discusses further development steps.
|
|
|
Fuzzy Neural Networks for Pattern Classification
Ollé, Tamás ; Kontár, Stanislav (oponent) ; Vágnerová, Jitka (vedoucí práce)
This work describes the principle of operation of neurons and how they form artificial neural networks. The structure and the operation of neurons are thoroughly described and the most widely used algorithm for neuron training is shown as well as the basics of fuzzy logic including its advantages and disadvantages. This work fully describes the backpropagation algorithm and the adaptive neuro-fuzzy inference system. These techniques provide effective methods of neural network learning.
|
|
|
Fuzzy Neural Networks for Pattern Classification
Ollé, Tamás ; Raida, Zbyněk (oponent) ; Vágnerová, Jitka (vedoucí práce)
This work describes the principle of operation of neurons and how they form artificial neural networks. The structure and the operation of neurons are thoroughly described and the most widely used algorithm for neuron training is shown as well as the basics of fuzzy logic including its advantages and disadvantages. This work fully describes the backpropagation algorithm and the adaptive neuro-fuzzy inference system. These techniques provide effective methods of neural network learning.
|
|
|
Parallelism in Digital Signal Processing
Mego, Roman ; Wyrzykowski, Roman (oponent) ; Vágnerová, Jitka (oponent) ; Frýza, Tomáš (vedoucí práce)
The doctoral thesis is focused on the systems for digital signal processing, its architecture and possibilities of software development. The text discussed the basic classification of computer systems from the view of parallel processing. It also demonstrates the behavior of the low-level and high-level programming languages on the multicore digital signal processors based on VLIW architecture. The aim of the dissertation thesis is to develop a tool that can be used to implement any DSP algorithm on the any VLIW processor with efficiency of the low-level programming languages, but with the advantages of the highlevel programming languages. Result is the software that uses a signal-flow graph approach to describe an algorithm, and generates the low-level assembly code.
|
|
|
LHC Abort Gap Monitor acquisition system
Pacner, Petr ; Vágnerová, Jitka (oponent) ; Kubíček, Michal (vedoucí práce)
The Beam Dump System of the LHC Large Hadron Collider is one of its critical systems. It is responsible for changing the trajectory of the beam towards the beam dump. For its functionality it uses an Abort Gap - a space with no particles within a beam structure. This gap is formed during the injection of the particles into the LHC. The injection is never ideal hence Abort Gap always contains some particles. Energy accumulated in the Abort Gap can cause various damages and so its amount has to be continuously estimated. This is the purpose of an Abort Gap Monitor. The currently deployed Abort Gap Monitor uses analog integrators to measure the particle population. This makes the instrument difficult to calibrate due to its hardware inaccuracies. To improve the system measurement versatility and reproducibility the effort is made to create an acquisition system with digital integration. This work aims to design, verify and implement the digital acquisition system. The digital acquisition system serves as one of the major parts of the new CERN SY-BI group Abort Gap Monitor. This document is split into four sections. The first briefly describes the structure of the Beam Dump System and the importance of the Abort Gap. The second section explains the principle of Abort Gap measurements and the data evaluation methodology. The third section describes the current system design and its issues. Then it focuses on the new Abort Gap Monitor topology followed by acquisition system development. The last section presents the simulations and measurements obtained with the new Abort Gap Monitor Acquisition System and discusses further development steps.
|
|
|
Fuzzy Neural Networks for Pattern Classification
Ollé, Tamás ; Kontár, Stanislav (oponent) ; Vágnerová, Jitka (vedoucí práce)
This work describes the principle of operation of neurons and how they form artificial neural networks. The structure and the operation of neurons are thoroughly described and the most widely used algorithm for neuron training is shown as well as the basics of fuzzy logic including its advantages and disadvantages. This work fully describes the backpropagation algorithm and the adaptive neuro-fuzzy inference system. These techniques provide effective methods of neural network learning.
|
|
|
Parallel Computing and Neural Networks in Behavioral Modeling
Vágnerová, Jitka ; Škvor,, Zbyněk (oponent) ; Dědková, Jarmila (oponent) ; Lukeš, Zbyněk (vedoucí práce)
This thesis is focused on methods for the aircraft equipment modeling. The first part provides a brief overview of classical system modeling approaches used for system description, identification, and modeling. Then adaptive, fuzzy and hybrid methods used mainly for black-box system modeling are introduced. Aim of the thesis is to develop an algorithm for identification and modeling of a general system, which can be nonlinear, dynamic and complex. Multiple inputs and multiple outputs of model are assumed. The main part of the thesis introduces a new method which falls into the hybrid systems. It combines fuzzy approach with parametrically defined rules and general regression neural network. Firstly, the fundamentals of simple general regression neural network and its smoothness parameter determination are presented. Secondly, the general regression neural network with the fuzzy rules is introduced. Third part of the thesis is focused on the parallel computing, one of the main objectives. The final algorithm is designed for the parallel machine, because the computational time can be significantly high and for the larger datasets, the model is not achievable when evaluated in single thread. Block diagram for parallel computing in Matlab and CUDA is provided, as well as the basic structure of CUDA source code. Finally, the method is verified on data obtained from the measurement of a miniaturized aircraft model using the antenna outside the aircraft and the probe inside the fuselage of the aircraft model. The validation of the method is done using mean squared error and compared to mean squared error of corresponding model performed using the multilayer neural network with backpropagation learning and Levenberg-Marquardt algorithm.
|
|
|
Fuzzy Neural Networks for Pattern Classification
Ollé, Tamás ; Raida, Zbyněk (oponent) ; Vágnerová, Jitka (vedoucí práce)
This work describes the principle of operation of neurons and how they form artificial neural networks. The structure and the operation of neurons are thoroughly described and the most widely used algorithm for neuron training is shown as well as the basics of fuzzy logic including its advantages and disadvantages. This work fully describes the backpropagation algorithm and the adaptive neuro-fuzzy inference system. These techniques provide effective methods of neural network learning.
|
host ::
RSS.