|
|
Learning dictionaries for sparse signal representation
Ettl, Ondřej ; Jirgl, Miroslav (referee) ; Mihálik, Ondrej (advisor)
This thesis focuses on examining data from measurements on a pressure mattress that uses a 30 × 11. sensor grid to collect. This data will be used to train and test learners dictionaries that are built using sparse signal representation. Applied learning methods include the method of optimal direction (MOD) and K-SVD, which uses singular decomposition. The resulting dictionaries for different numbers of iterations or atoms are then used to classify and reconstruction of the test data. Cross-validation determined the true positive ratio of the models, which was then compared with conventional classifiers. These models included Decision trees, KNNs and SVMs. Finally, the ability of the learning dictionaries was verified to filter the error in the corrupted image.
|
|
|
Multiloop dynamical models of human driver
Čelko, Jakub ; Jirgl, Miroslav (referee) ; Mihálik, Ondrej (advisor)
The thesis is focused on modelling human-machine systems, where a human operator controls a vehicle. The result of the work is a comparison of the prediction capabilities of chosen multiloop and single-loop human driver models. These models are applied in scenarios where, in addition to the lane change forcing function, there is also a disturbance signal that affects the vehicle’s motion. The parameters characterising the human dynamics are extracted from the model of the whole loop, defined as a discrete state-space model. The thesis also contains research on driver models, description of the used vehicle simulator and the Unreal Engine 4 environment used for its implementation. Documentation of implemented disturbance signals that act as a fault at the input of the regulated system is also presented. The identification results on the measured data suggest that the models relying on a single transfer function are losing prediction capabilities in scenarios, where the disturbances affect the vehicle. In such cases, the multiloop models with the same number of parameters are able to capture human behaviour more accurately.
|
|
|
Accelerometer-based monitoring of steering wheel angle
Navrátil, Tadeáš ; Jirgl, Miroslav (referee) ; Mihálik, Ondrej (advisor)
This thesis deals with the design of a device which will be used for measuring the steering wheel angle. Main goal was to design basic concept of device which is using data gathered from three accelerometers in the reconstruction algorithm to get steering wheel angle. Discussion of various reconstruction algorithms (with regards to simulated output noise levels from acceleromater) and final Implementation of reconstruction algorithm (in MATLAB simulink enviroment) is also part of this thesis. Furthermore, thesis deals with the selection of suitable components. Mainly selection of accelerometers, microcontroller and the selection of a suitable communication bus for mutual communication between them. Thesis also deals with design improvements of algorithm and realization of the whole device. In the end, we compare results of measuring device and inner senzor of steering wheel.
|
|
|
Analysis of car drivers’ behaviour and driving style
Novik, Svatoslav ; Jirgl, Miroslav (referee) ; Mihálik, Ondrej (advisor)
Driving safety remains one of the important issues. Various assistance systems are also implemented, which focus mainly on the analysis of driving and control of the car by the driver. To understand the performance of the driver’s control, a program was created to obtain valuable data and relevant characteristics. To obtain the data, we used an internally designed, laboratory-made vehicle driving simulator. Driver data was measured from the motor vehicle driving simulation, and these were examined and evaluated in the MATLAB environment thanks to integral criteria and other calculated parameters. The obtained values were then used as a training set for the machine learning, using LDA and QDA methods (linear and quadratic discriminant analyzing). This method then provided information on which values are most appropriate and then determines as precisely as possible which driver may be involved in future measures, when this information is not stated in advance.
|
|
|
Bandlimited signals, their properties and extrapolation capabilities
Mihálik, Ondrej ; Havránek, Zdeněk (referee) ; Jura, Pavel (advisor)
The work is concerned with the band-limited signal extrapolation using truncated series of prolate spheroidal wave function. Our aim is to investigate the extent to which it is possible to extrapolate signal from its samples taken in a finite interval. It is often believed that this extrapolation method depends on computing definite integrals. We show an alternative approach by using the least squares method and we compare it with the methods of numerical integration. We also consider their performance in the presence of noise and the possibility of using these algorithms for real-time data processing. Finally all proposed algorithms are tested using real data from a microphone array, so that their performance can be compared.
|
|
|
Application of Legendre basis for spectral analysis
Mesárošová, Michaela ; Jirgl, Miroslav (referee) ; Mihálik, Ondrej (advisor)
The thesis focuses on the possibilities of using Legendre polynomials in order to obtain a spectrum of signals. It examines their properties in the time and frequency domain such as generating methods, root position, and orthogonality. Another goal was to implement the Legendre transform and to verify the quality of the obtained spectra and signal approximations in comparison with various methods. Finally, it deals with the choice of a suitable approximation order as well as the analytical possibilities of spectrum calculation.
|
|
|
Modelling of a heat water exchanger and its regulation
Černík, Michal ; Kaczmarczyk, Václav (referee) ; Mihálik, Ondrej (advisor)
The thesis deals with the modelling of a hot water heat exchanger station and its control. It includes project documentation, with theoretical and practical part. The theoretical part is dealing with theoretical knowledge about every used element that is used in the design. The practical part deals with the implementation of the electrical wiring of the station, its connection with the control automaton, design of the mathematical model, simulation and design of the controller.
|
|
|
Application for Analysis of McRuer's Models Parameters
Pohorský, Ondřej ; Mihálik, Ondrej (referee) ; Jirgl, Miroslav (advisor)
This bachelor’s thesis deals with design and implementation of Application for automatic analysis parameters of McRuer’s models. This work contains an explanation of McRuer’s models based on Crossover law theory. Next part is description of System Identification toolbox and description of how to get model from measured data. In this work you can find information about data acquisition for human model estimation. Application is designed in MATLAB App designer, which is described in more detail in this work.
|
|
|
McRuer Models for Human-Machine Systems
Čelko, Jakub ; Jirgl, Miroslav (referee) ; Mihálik, Ondrej (advisor)
Bachelor’s thesis is focused on modelling of the human-machine systems using McRuer models, particularly systems formed by human operator and a steering wheel. The aim of the work is to design and measure the responses of the human operator in the regulation of different types of systems and different system forcing functions. Different steering wheel stiffness was also measured to find its influence on the speed and quality of the regulation. Results of this work are mathematical descriptions of the operator for the individual scenarios. The descriptions are expressed as transfer functions, with the operator being considered as linear time-invariant system.
|
|
|
Modelling and Assessment of Driver Behaviour Using Fuzzy Logic
Radvanský, Martin ; Mihálik, Ondrej (referee) ; Jirgl, Miroslav (advisor)
The work contains a research of mathematical modeling with a focus on the driver's behavior while driving. The following text contains the basic statistical processing of data obtained from the driving simulator for a total of 38 drivers. Subsequently, possible approaches to the use of fuzzy systems for modeling driver behavior are proposed. The main part of the work is focused on the verification of three hypotheses related to the use of fuzzy inference systems to model driver behavior.
|
guest ::
RSS feed.