# Fakulta elektrotechniky a informatiky / Faculty of Electrical Engineering and Informatics

Stálý URI pro tuto komunituhttps://hdl.handle.net/10195/3847

Práce obhájené před rokem 2008 jsou uloženy pouze v kolekci Vysokoškolské kvalifikační práce

## Procházet

38 výsledky

## Search Results

Konferenční objektpeer-reviewedpostprint Omezený přístup Passage Detection of a Train via a Reference Point(Springer, 2023) Rejfek, Luboš; Pidanič, Jan; Štursa, Dominik; Nguyen, Tan N.; Tran, Phuong T.; Němec, Zdeněk; Zálabský, TomášA reference point detection system for position validation of a mobile object was developed for verification of experiments. The detection is based on a classic image processing algorithm and a processing algorithm using neural networks. Both approaches are compared. High-precision concept of the system is based on a camera sensor and automatic processing of video frames for position evalua-tion. The designed system was tested on a real application proving correct operation.Konferenční objektpeer-reviewedpostprint Omezený přístup Resistance of the radio modems against the narrow band interference(IEEE (Institute of Electrical and Electronics Engineers), 2023) Rejfek, Luboš; Rozsíval, Pavel; Zálabský, Tomáš; Nguyen, Tan N; Tran, Phuong T; Stejskal, VojtechThis electronic document is aimed at the influence of interference signals on a communication link. The theory was tested on a real system. The communication link was realized by radio modems RipEx. These modems are optimal devices for the realization of the experiment because they can use different settings for every modulation. One of four modulations can be chosen for the creation of the wireless PC network between the moving systems in these radio modems. The tests were realized for all possible modulations and different channel bandwidths. The modulations used were based on frequency shift keying and quadrature amplitude modulation. The results from the comparison are described in the conclusion.Článekpeer-reviewedpostprint Omezený přístup Whitening Filters Application for Ionospheric Propagation Delay Extraction(IEEE (Institute of Electrical and Electronics Engineers), 2023) Rejfek, Luboš; Juryca, Karel; Nguyen, Tan N. N; Beran, Ladislav; Voznak, MiroslavPreprocessed signals from Global Positioning System (GPS) satellites can be used for ionospheric research. Important ionospheric characteristics are for example ionospheric propagation delays and scintillations. These characteristics are important for understanding the ionosphere and showing how the ionosphere reacts to external forces (e.g., geomagnetic disturbance, solar eclipse, etc.). Dual-frequency systems use an approach based on the difference between two pseudoranges. However, this approach has two possible disadvantages: 1) the price of the dual frequency receiver is relatively high and 2) both pseudoranges can affect the obtained results so the operators cannot separately analyze their relative contribution. The objective of the article is to demonstrate an algorithm for processing data from a GPS receiver. The algorithm does not merge the pseudoranges obtained from different carrier waves. Accordingly, it was decided to use whitening filters to accomplish this task for every frequency separately. Compared with the dual-frequency approach, the obtained results of the newly developed algorithm demonstrate that this algorithm works accurately, and the propagation delays can be estimated independently of the used frequency.Konferenční objektpeer-reviewedpostprint Omezený přístup FMICW Radar Target Classification By Neural Network(IEEE (Institute of Electrical and Electronics Engineers), 2020) Pitaš, Karel; Rejfek, Luboš; Nguyen, Tan N.; Beran, Ladislav; Tran, Phuong T.; Fišer, OndřejThis document describes automatic classification of targets detected by the FMICW radar. These targets are counted and sorted to three groups (incoming, outgoing and static targets). We derived this information from the output of the neural network which marked the targets in 2D spectrum. The additional neural network has five layers. The first layer is used for the suppression of the targets with even numbers of points, which causes problems during the symmetry detection. The second and third layers detect the symmetry in the dimension (vertical or horizontal). The fourth layer checks out if the symmetry is in both dimensions and if the detection is not a false alert caused by the constellation of the targets. The fifth layer contains only 4 neurons and this layer is used for counting of the targets and classification of the targets (if they are static, incoming or outgoing). The neural network is composed of a simple block for the easy implementation on the FPGA.Článekpeer-reviewedpostprint Otevřený přístup Speaker Verification Using Autoregressive Spectrum of Speech Signal in Composite Vector Stochastic Processes Model Representation(Institute of Mechanics of Continua and Mathematical Sciences, 2019) Chmelařová, Natalija; Tykhonov, Vyacheslav A.; Bezruk, Valerij M.; Chmelař, Pavel; Rejfek, LubošThis paper deals with the speaker verification system similar to a fingerprint or an eye scanner. For these purpose a long-term words' model and its spectral characteristics were used. The speaker verification method uses the word's sound parametric spectrum factorization in composite vector stochastic process representation based on the multiplicative autoregressive model. The developed method enables to receive the words' features with stable characteristics for the same speaker and differ for different speakers. During the training phase speaker's etalon frequencies has to be estimated for a pronounced word repeated several times. In the verification phase a speaker pronouncing the same word, word's frequencies are estimated and compared with the etalon frequencies database to find the best match or his deny. The results presented in the paper showed the high correct identification probability.Konferenční objektpeer-reviewedpostprint Otevřený přístup Nonlinear distortion in a microwave high power amplifier(IEEE (Institute of Electrical and Electronics Engineers), 2019) Bezoušek, Pavel; Matoušek, David; Rejfek, LubošIn this paper the high-power amplifier CGHV31500F designed for the radar S-Band application is studied. The amplifier module is based on GaN HEMT encapsulated internally matched transistor, delivering more than 500 W in a pulse regime. Nonlinear model of this amplifier was developed and nonlinear distortion of a standard radar signal and of possible future QAM signals are predicted.Konferenční objektpeer-reviewedpostprint Otevřený přístup Total Electron Content Measurements by Single-Frequency GPS Receiver(IEEE (Institute of Electrical and Electronics Engineers), 2019) Rejfek, Luboš; Urbar, Jaroslav; Pitaš, Karel; Chmelař, PavelIn this paper is described the measurement of the Total Electron Content by the single-frequency GPS receiver. Disadvantage of this method is that it cannot be used for real-time processing. The advantages of the proposed method are that the device is cheaper than dual-frequency GPS receiver and it can reinterpret older valuable measurements by such single-frequency receivers. The algorithm was tested with the dual-frequency receiver. Signal from selected frequency (f1, or f2) was processed by presented algorithm and results were compared with standard processing of the signals, which is used for the dual-frequency receivers. Results are promising.Konferenční objektpeer-reviewedpostprint Otevřený přístup The Fine Plane Range Estimation From Point Cloud(IEEE (Institute of Electrical and Electronics Engineers), 2019) Chmelařová, Natalija; Chmelař, Pavel; Rejfek, LubošThis paper deals with the fine plane estimation. For a plane detection the level connected component labeling algorithm is used. The precise plane position is determined by the plane's points histogram in the scanning dimension. Up till now, the final plane's range has been selected manually by a parameter. Plane presence in an input point cloud represents the higher points' concentration in the scanning dimension, higher than average value. The concentration decreases on both sides from a local maximum value. This paper presents the algorithm of the elbows' position estimation for the optimal plane's points range expression. The experimental part shows practical results of several examples and gives recommendations for parameters setting.Konferenční objektpeer-reviewedpostprint Otevřený přístup The Automatic Undistortion Strength Estimation for Any Describable Optical Distortion(IEEE (Institute of Electrical and Electronics Engineers), 2019) Chmelařová, Natalija; Chmelař, Pavel; Rejfek, LubošA presence of an optical distortion does undesired influence on the geometrical image quality. The main reason of its occurring is using a lens with low or variable focal length or simple camera optics. When an image captured by a camera is affected by undesired distortion and it is used for the distance measurement, it can cause significant measurement errors. The aim of this paper is to introduce a method for finding an optimal undistortion parameter for any optical distortion, which can be described mathematically. The proposed method uses the Hough transformation accumulator and at each iteration step the quality of straight lines is evaluated as a gradient from local maximum. The algorithm is verified and assessed on several images. The developed approach can be used to find the global maximum gradient in general, even when a local maximum changes its size over iterations.Článekpeer-reviewedpublished version Otevřený přístup Neural Networks Application for Processing of the Data from the FMICW Radars(MDPI, 2019) Rejfek, Luboš; Nguyen, Tan N.; Chmelař, Pavel; Beran, Ladislav; Phuong, Tran T.In this paper the results of the Neural Networks and machine learning applications for radar signal processing are presented. The radar output from the primary radar signal processing is represented as a 2D image composed from echoes of the targets and noise background. The Frequency Modulated Interrupted ContinuousWave (FMICW) radar PCDR35 (Portable Cloud Doppler Radar at the frequency 35.4 GHz) was used. Presently, the processing is realized via a National Instruments industrial computer. The neural network of the proposed system is using four or five (optional for the user) signal processing steps. These steps are 2D spectrum filtration, thresholding, unification of the target, target area transforming to the rectangular shape (optional step), and target board line detection. The proposed neural network was tested with sets of four cases (100 tests for every case). This neural network provides image processing of the 2D spectrum. The results obtained from this new system are much better than the results of our previous algorithm.