Driver sleepiness is a cause for crashes and it is estimated that 3.9 to 33 % of all crashes might be related to sleepiness at the wheel. It is desirable to get an objective measurement of driver sleepiness for reduced sensitivity to subjective variations. Using deep learning for classification of driver sleepiness could be a step toward this objective. In this master thesis, deep learning was used for investigating classification of electrophysiological data, electroencephalogram (EEG) and electrooculogram (EOG), from drivers into levels of sleepiness. The EOG reflects eye position and EEG reflects brain activity. Initially, the intention was to include electrocardiogram (ECG), which reflects heart activity, in the research but this data were later excluded. Both raw time series data and data transformed into time-frequency domain representations were fed into the developed neural networks and for comparison manually extracted features were used in a shallow neural network architecture. Investigation of using EOG and EEG data separately as input was performed as well as a combination as input. The data were labeled using the Karolinska Sleepiness Scale, and the scale was divided into two labels "fatigue" and "alert" for binary classification or in five labels for comparison of classification and regression. The effect of example length was investigated using 150 seconds, 60 seconds and 30 seconds data. Different variations of the main network architecture were used depending on the data representation and the best result was given when using a combination of a convolutional neural network (CNN) and a long short-term memory (LSTM) network with time distributed 150 seconds EOG data as input. The accuracy was in this case 80.4 % and the majority of both alert and fatigue epochs were classified correctly with 85.7 % and 66.7 % respectively. Using the optimal threshold from the created receiver operating characteristics (ROC) curve resulted in a more balanced classifier with 76.3 % correctly classified alert examples and 79.2 % correctly classified fatigue examples. The results from the EEG data, both in terms of accuracy and distribution of correctly classified examples, were shown to be less promising compared to EOG data. Combining EOG and EEG signals was shown to slightly increase the proportion of correctly classified fatigue examples. However, more promising results were obtained when balancing the classifier for solely EOG signals. The overall result from this project shows that there are patterns in the data connected to sleepiness that the neural network can find which makes further work on applying deep learning to the area of driver sleepiness interesting.

The thesis describes hardware and software of Control system for cleaning station CIP used for cleaning of milk processing equipment. The description of hardware contains definitions and number of PLC modules for PLC Simatic S7-300. The software part contains programs for PLC, SCADA software for WinCC flexible 2005 and the description of the computer's model.

One of the most common methods for isolated words recognition is based on Hidden Markov models. Speech signal can be considered as a sequence of successive parts of the signal with specific statistical parameters. Hidden Markov model corresponds to the statistical model with the final number of states, which may be useful for signals such as speech. HTK module is a software tools, which is mostly used to work with hidden Markov models.

The aim of this thesis is creation the appropriate multimedia support for signals and system with continuous time. The understanding of this issue is very important, because the obligatory subject Signals and systems, exactly BSIS, is taught at the EST bachelor degree. The understanding is also necessary prerequisite to successful understanding next topics in other related subjects. The next part of this thesis is focused on one dimension discrete signals. Concretely, the aim of this part is a realization of software system. Designed system has some basic operations (the signal energy, the number of signal zero crossing etc.) with sound files and also some advance functions e.g. vowel seeking and separating in fluent speech. The system is divided into two main parts. The first one analyzes sound files, creates the new sound file with wanted vowel and matrices with important parameters for other processing. The second program computes with given data, which statistically evaluates in other steps. The final system can be useful for speaker recognition, his emotional status etc.

This master thesis describes the principles of the Global Navigation Satellite System GNSS, specifically the GPS, GLONASS and Galileo systems. The thesis analyzes the structure of individual GNSS subsystems and introduces their properties. The algorithm for calculating the position is designed in the interactive programming environment MATLAB for the processing of GPS and GLONASS sampled signals. The position is calculated by a distance measurement method which searches for the intersection of spherical surfaces. The calculation is designed for four satellites and when more satellites are detected, the calculation is repeated for all possible combinations. From this position the combination with the lowest DOP (Dilution of Precision) factor is determined, and the calculation of the position is repeated for the best constellation of satellites. In this thesis the user graphical interface for entering the input of data, input parameters and the display of calculated values are created. Finally the calculation of the measured data is displayed on the selected location online map portal

This diploma thesis deals with implementation of fast serial bus and SATA controler in the FPGA chip. The work is divided into two parts. In the first part the circuit for communication between the FPGAs is designed and in the second part the circuit for direct connection of SATA hard disk to a gate array is created. The circuit for communication between the FPGA is designed according to SATA specification. Link layer and physical layers are implemented in VHDL with programmable logic resources.

This master’s thesis deals with the problems of a technical analysis and its use. The first part of thesis describes theoretical background of the technical analysis and basic concepts and principles of the currency market Forex. The second part is devoted to analyzing the current situation in the environment of currency market. The output of the thesis is a desktop application for the support of technical analysis. The design and development of the application is described in the last part of this thesis.

De nos jours, il existe de nombreuses applications liées à la vision et à l’audition visant à reproduire par des machines les capacités humaines. Notre intérêt pour ce sujet vient du fait que ces problèmes sont principalement modélisés par la classification de signaux temporels. En fait, nous nous sommes intéressés à deux cas distincts, la reconnaissance de la démarche humaine et la reconnaissance de signaux audio, (notamment environnementaux et musicaux). Dans le cadre de la reconnaissance de la démarche, nous avons proposé une nouvelle méthode qui apprend et sélectionne automatiquement les parties dynamiques du corps humain. Ceci permet de résoudre le problème des variations intra-classe de façon dynamique; les méthodes à l’état de l’art se basant au contraire sur des connaissances a priori. Dans le cadre de la reconnaissance audio, aucune représentation de caractéristiques conventionnelle n’a montré sa capacité à s’attaquer indifféremment à des problèmes de reconnaissance d’environnement ou de musique : diverses caractéristiques ont été introduites pour résoudre chaque tâche spécifiquement. Nous proposons ici un cadre général qui effectue la classification des signaux audio grâce à un problème d’apprentissage de dictionnaire supervisé visant à minimiser et maximiser les variations intra-classe et inter-classe respectivement. / Nowadays, there are a lot of applications related to machine vision and hearing which tried to reproduce human capabilities on machines. These problems are mainly amenable to a temporal signals classification problem, due our interest to this subject. In fact, we were interested to two distinct problems, humain gait recognition and audio signal recognition including both environmental and music ones. In the former, we have proposed a novel method to automatically learn and select the dynamic human body-parts to tackle the problem intra-class variations contrary to state-of-art methods which relied on predefined knowledge. To achieve it a group fused lasso algorithm is applied to segment the human body into parts with coherent motion value across the subjects. In the latter, while no conventional feature representation showed its ability to tackle both environmental and music problems, we propose to model audio classification as a supervised dictionary learning problem. This is done by learning a dictionary per class and encouraging the dissimilarity between the dictionaries by penalizing their pair- wise similarities. In addition the coefficients of a signal representation over these dictionaries is sought as sparse as possible. The experimental evaluations provide performing and encouraging results.

<p>Acoustic signals have long been used to monitor the performance of machinery containing mechanical moving parts, especially machinery used in manufacturing. Rotating components generate harmonic signals with a fundamental frequency corresponding to the period of rotation, although the fundamental frequency and some of the harmonics may be missing. In addition, the meshing of the teeth in gears generates higher frequencies corresponding to the period of the gear teeth interaction. We call the former frequencies harmonic frequencies and the latter frequencies strong tone frequencies. Each strong tone frequency typically has associated with it, a set of modulation frequencies.<br></p><p><br></p><p>For each strong tone frequency, it is important to be able to determine which modulation frequencies correspond to a particular harmonic series, since this can help to determine which component in the overall mechanism is failing. In this work, we describe a novel process for selecting from a set of candidate modulation frequencies that comprise one or more harmonic sequences.<br></p><p><br></p><p>We also describe the signal processing pipeline used to extract the frequency components from the raw acoustic signal.<br></p>

<div>Surface soil moisture is one of the few direct hydrological variables which can be measured. It plays a crucial part in the water cycle, agriculture, drought development, runoff generation, and many other phenomena. Satellite observations from active and passive microwave radiometers are best suited for the retrieval of soil moisture. The relationship between soil dielectric constant and water content is direct and is used to determine the surface soil moisture levels. Active microwave remote sensing techniques measure the energy reflected from target surfaces (ocean, soil, biomass) after transmitting a pulse of microwave energy, whereas passive microwave sensors measure the self-emissions of the target surfaces. The passive missions by ESA's SMOS and NASA's SMAP have demonstrated this technology for remote sensing on a global scale. Global Navigation Satellite System-Reflectometry (GNSS-R) is an alternative approach to the remote sensing of soil moisture, as demonstrated through several airborne and ground-based experiments. The new technique of Signals of Opportunity (SoOp) uses a bistatic radar configuration in which the non-cooperative transmitter already transmits signals designed for communication or navigation. The receiver reuses the reflected energy from the target surface (ocean, soil, biomass), thereby making the digital communication and navigation signal spectrum useful to the remote sensing science community. Several airborne and ground-based experiments have been conducted on the use of digital communication signals, a range of frequencies from P-band to Ku-band, for measurement of ocean surface roughness, wind speed, and soil moisture. </div><div> </div><div>This thesis presents the preliminary results obtained for reflectivity retrievals for the 2017 and 2018 S-band tower-based SoOp field experiment conducted at Purdue's Agronomy Center for Research and Education (ACRE). XM signals were recorded by a sky-facing antenna and an Earth-facing antenna mounted atop a tower. The line-of-sight (direct) signal is captured by the sky-facing antenna and reflected signal from the soil captured by the Earth-facing antenna was used for the ambiguity function of XM transmission. A link budget was used to determine the received signal to noise ratio (SNR). The cross-correlation between the direct and the reflected XM signals was used to estimate reflectivity values. The reflectivity retrievals were compared with the in-situ soil moisture content at 5 cm depth obtained by the HydraProbes. The reflectivity values were also verified by a Signals of Opportunity (SoOp) Coherent Bistatic (SCoBi) simulated model.</div>