Práce se zabývá návrhem a implementací testovacího zařízení pro manipulačnínmechanismus vzorku v elektronovém mikroskopu. Testovací zařízení a jeho software zajištuje meření několika parametrů mechanismu, jejich statistické vyhodnocení a porovnání se specifikací. Cílem je vytvořit komplexní testovací zařízení s jednoduchým uživatelským rozhraním, s požadavkem náhrady stávajícího nemodulárního a nestabilního řešení a jeho rozšíření o testování dalších parametrů.

This diploma thesis deals with the design and control of the mobile robot with caterpillar tracks on the basis of information obtained via infrared thermocouple. The work includes firmware implementation for robot chassis control and communication with the sensors used. The functionality of the proposed system is demonstrated on a robotic task when the robot is searching the source of excessive heat within his surroundings. The theoretical part deals with the basic description of the common robotic chassis, contactless temperature measurement, further analysis of the components used and their principles. The practical part is devoted to the firmware implementation and detailed description of selected problems. In the end of the thesis there is a summarization of the achievements and the success of the robot while detection of the suspicious objects with high surface temperature.

This thesis describes circuit design and realization for differential control of mobile robot. Control system is based on microcontroller STM32 of company STMicroelectronics, which selection is discussed too. Part of the work thesis is description issues concerning differential control of mobile robot, DC motor cotrol with incremental encoder including cascade control and implementation of cotrol algorithms.

Solicitud de envío manuscrito de artículo científico. / El presente trabajo propone 4 alternativas de codificadores inspirados en el codificador MPEG-1, capa I, descrito en el estándar ISO/IEC 11172-3. El problema que se intenta resolver es el de requerir definir un modelo psicoacústico explícitamente para lograr codificar audio, reemplazándolo por redes neuronales. Todas las alternativas de codificador están basadas en redes neuronales convolucionales multiescala (MCNN) que emulan el modelo psicoacústico 1 del codificador mencionado. Las redes tienen 32 entradas que corresponden a las 32 subbandas del nivel de presión sonora (SPL – sound pressure level), y una única salida que corresponde a una de las 32 subbandas de o bien la relación señal a máscara (SMR) o bien el vector de asignación de bits. Es decir, un codificador está compuesto de un conjunto de 32 redes neuronales. La validación empleó los 10 primeros segundos de 15 canciones elegidas aleatoriamente de 10 géneros musicales distintos. Se comparó la calidad de las señales de audio generadas por cada codificador contra la de MPEG-1, capa I, mediante la métrica de ODG. El codificador cuya entrada es el SPL y cuya salida es la SMR, planteado por Guillermo Kemper, obtuvo los mejores resultados al realizar la comparación para 96 kbps y 192 kbps. El codificador denominado “SBU1” obtuvo los mejores resultados para 128 kbps. / The present work proposes 4 encoder alternatives, inspired in the MPEG-1, layer I encoder described in the ISO/IEC 11172-3 standard. The problem addressed here is the requirement of explicitly defining a psychoacoustic model to code audio, instead replacing it by neural networks. All the proposals are based on multiscale convolutional neural networks (MCNN) that emulate the psychoacoustic model 1 of the referred encoder. The networks have 32 inputs that map the 32 subbands of the sound pressure level (SPL), and a single output that corresponds to each of the 32 subbands of either the signal-to-mask ratio (SMR) or the bit allocation vector. Thus, an encoder is composed of a set of 32 neural networks. The validation process took the first 10 seconds of 15 randomly chosen songs of 10 different musical genres. The audio signal quality of the proposed encoders was compared to that of the MPEG-1, layer I encoder, using the ODG metric. The encoder whose input is the SPL and whose output is the SMR, proposed by Guillermo Kemper, yielded the best results for 96 kbps and 192 kbps. The encoder named “SBU1” had the best results for 128 kbps. / Tesis

With an increasing amount of information on the internet, automatic text summarization could potentially make content more readily available for a larger variety of people. Training and evaluating text summarization models require datasets of sufficient size and quality. Today, most such datasets are in English, and for minor languages such as Swedish, it is not easy to obtain corresponding datasets with handwritten summaries. This thesis proposes methods for compiling high-quality datasets suitable for abstractive summarization from a large amount of noisy data through characterization and filtering. The data used consists of Swedish news articles and their preambles which are here used as summaries. Different filtering techniques are applied, yielding five different datasets. Furthermore, summarization models are implemented by warm-starting an encoder-decoder model with BERT checkpoints and fine-tuning it on the different datasets. The fine-tuned models are evaluated with ROUGE metrics and BERTScore. All models achieve significantly better results when evaluated on filtered test data than when evaluated on unfiltered test data. Moreover, models trained on the most filtered dataset with the smallest size achieves the best results on the filtered test data. The trade-off between dataset size and quality and other methodological implications of the data characterization, the filtering and the model implementation are discussed, leading to suggestions for future research.

Background: In Autonomous Driving Vehicles, the vehicle receives pixel-wise sensor data from RGB cameras, point-wise depth information from the cameras, and sensors data as input. The computer present inside the Autonomous Driving vehicle processes the input data and provides the desired output, such as steering angle, torque, and brake. To make an accurate decision by the vehicle, the computer inside the vehicle should be completely aware of its surroundings and understand each pixel in the driving scene. Semantic Segmentation is the task of assigning a class label (Such as Car, Road, Pedestrian, or Sky) to each pixel in the given image. So, a better performing Semantic Segmentation algorithm will contribute to the advancement of the Autonomous Driving field. Research Gap: Traditional methods, such as handcrafted features and feature extraction methods, were mainly used to solve Semantic Segmentation. Since the rise of deep learning, most of the works are using deep learning to dealing with Semantic Segmentation. The most commonly used neural network architecture to deal with Semantic Segmentation was the Convolutional Neural Network (CNN). Even though some works made use of Recurrent Neural Network (RNN), the effect of RNN in dealing with Semantic Segmentation was not yet thoroughly studied. Our study addresses this research gap. Idea: After going through the existing literature, we came up with the idea of “Using RNNs as an add-on module, to augment the skip-connections in Semantic Segmentation Networks through residual connections.” Objectives and Method: The main objective of our work is to improve the Semantic Segmentation network’s performance by using RNNs. The Experiment was chosen as a methodology to conduct our study. In our work, We proposed three novel architectures called UR-Net, UAR-Net, and DLR-Net by implementing our idea to the existing networks U-Net, Attention U-Net, and DeepLabV3+ respectively. Results and Findings: We empirically showed that our proposed architectures have shown improvement in efficiently segmenting the edges and boundaries. Through our study, we found that there is a trade-off between using RNNs and Inference time of the model. Suppose we use RNNs to improve the performance of Semantic Segmentation Networks. In that case, we need to trade off some extra seconds during the inference of the model. Conclusion: Our findings will not contribute to the Autonomous driving field, where we need better performance in real-time. But, our findings will contribute to the advancement of Bio-medical Image segmentation, where doctors can trade-off those extra seconds during inference for better performance.

Wireless communications systems are used today in a variety of milieux, with a recurring theme: users and applications regularly require higher throughput. Multiple antennas enable higher throughput and/or more robust performance than single-antenna communications, with no increase in power or frequency bandwidth. Systems are required which achieve the full potential of this "space-time" communication channel under the significant challenges of time-varying fading, multiple users, and the choice of appropriate coding schemes. This dissertation is focused on solutions to these problems. For the single-user case, there are many well-known coding techniques available; in the first part of this dissertation, the performance of two of these methods are analyzed. Trained and differential modulation are simple coding techniques for single-user time-varying channels. The performance of these coding methods is characterized for a channel having a constant specular component plus a time-varying diffuse component. A first- order auto-regressive model is used to characterize diffuse channel coefficients that vary from symbol to symbol, and is shown to lead to an effective SNR that decreases with time. A lower bound on the capacity of trained modulation is found for the specular/diffuse channel. This bound is maximized over the training length, training frequency, training signal, and training power. Trained modulation is shown to have higher capacity than differential coding, despite the effective SNR penalty of trained modulation versus differential methods. The second part of the dissertation considers the multi-user, multi-antenna channel, for which capacity-approaching codes were previously unavailable. Precoding with the channel inverse is shown to provide capacity that approaches a constant as the number of users and antennas simultaneously increase. To overcome this limitation, a simple encoding algorithm is introduced that operates close to capacity at sum-rates of tens of bits/channel-use. The algorithm is a variation on channel inversion that regularizes the inverse and uses a "sphere encoder" to perturb the data to reduce the energy of the transmitted signal. Simulation results are presented which support our analysis and algorithm development.

Understanding how users interact with a company's service is essential for data-driven businesses that want to better cater to their users and improve their offering. By using a generative machine learning approach it is possible to model user behaviour and generate new data to simulate or recognize and explain typical usage patterns. In this work we introduce an approach for modelling users' interaction behaviour at scale in a client-service model. We propose a novel representation of multivariate time-series data as time pictures that express temporal correlations through spatial organization. This representation shares two key properties that convolutional networks have been built to exploit and allows us to develop an approach based on deep generative models that use convolutional networks as backbone. In introducing this approach of feature learning for time-series data, we expand the application of convolutional neural networks in the multivariate time-series domain, and specifically user interaction data. We adopt a variational approach inspired by the β-VAE framework in order to learn hidden factors that define different user behaviour patterns. We explore different values for the regularization parameter β and show that it is possible to construct a model that learns a latent representation of identifiable and different user behaviours. We show on real-world data that the model generates realistic samples, that capture the true population-level statistics of the interaction behaviour data, learns different user behaviours, and provides accurate imputations of missing data. / Förståelse för hur användare interagerar med ett företags tjänst är essentiell för data-drivna affärsverksamheter med ambitioner om att bättre tillgodose dess användare och att förbättra deras utbud. Generativ maskininlärning möjliggör modellering av användarbeteende och genererande av ny data i syfte att simulera eller identifiera och förklara typiska användarmönster. I detta arbete introducerar vi ett tillvägagångssätt för storskalig modellering av användarinteraktion i en klientservice-modell. Vi föreslår en ny representation av multivariat tidsseriedata i form av tidsbilder vilka representerar temporala korrelationer via spatial organisering. Denna representation delar två nyckelegenskaper som faltningsnätverk har utvecklats för att exploatera, vilket tillåter oss att utveckla ett tillvägagångssätt baserat på på djupa generativa modeller som bygger på faltningsnätverk. Genom att introducera detta tillvägagångssätt för tidsseriedata expanderar vi applicering av faltningsnätverk inom domänen för multivariat tidsserie, specifikt för användarinteraktionsdata. Vi använder ett tillvägagångssätt inspirerat av ramverket β-VAE i syfte att lära modellen gömda faktorer som definierar olika användarmönster. Vi utforskar olika värden för regulariseringsparametern β och visar att det är möjligt att konstruera en modell som lär sig en latent representation av identifierbara och multipla användarbeteenden. Vi visar med verklig data att modellen genererar realistiska exempel vilka i sin tur fångar statistiken på populationsnivå hos användarinteraktionsdatan, samt lär olika användarbeteenden och bidrar med precisa imputationer av saknad data.

An issue authorities and banks face is fraud related to payments and transactions where huge monetary losses occur to a party or where money laundering schemes are carried out. Previous work in the field of machine learning for fraud detection has addressed the issue as a supervised learning problem. In this thesis, we propose a model which can be used in a fraud detection system with transactions and payments that are unlabeled. The proposed modelis a Long Short-term Memory in an auto-encoder decoder network (LSTMAED)which is trained and tested on transformed data. The data is transformed by reducing it to Principal Components and clustering it with K-means. The model is trained to reconstruct the sequence with high accuracy. Our results indicate that the LSTM-AED performs better than a random sequence generating process in learning and reconstructing a sequence of payments. We also found that huge a loss of information occurs in the pre-processing stages. / Obehöriga transaktioner och bedrägerier i betalningar kan leda till stora ekonomiska förluster för banker och myndigheter. Inom maskininlärning har detta problem tidigare hanterats med hjälp av klassifierare via supervised learning. I detta examensarbete föreslår vi en modell som kan användas i ett system för att upptäcka bedrägerier. Modellen appliceras på omärkt data med många olika variabler. Modellen som används är en Long Short-term memory i en auto-encoder decoder nätverk. Datan transformeras med PCA och klustras med K-means. Modellen tränas till att rekonstruera en sekvens av betalningar med hög noggrannhet. Vår resultat visar att LSTM-AED presterar bättre än en modell som endast gissar nästa punkt i sekvensen. Resultatet visar också att mycket information i datan går förlorad när den förbehandlas och transformeras.

Asynchronous multiplexing schemes are efficient than synchronous schemes, in a bursty traffic environment of multiple access local area network (LAN), as fixed bandwidth is not allocated among the users and there is no access delay. Fiber- Optic Code-Division Multiple Access (FO-CDMA) is one such asynchronous multiplexing scheme suitable for high speed LAN networks. While FO-CDMA offers potential benefits it also faces challenges in three diverse areas which are 1) coding algorithms and schemes 2) advanced encoding and decoding hardware and 3) network architecture. In this thesis, as a solution to the first challenge, we propose the design and construction of a new family of codes, wavelength/time multiple-pulses-per-row (W/T MPR) codes. These codes have good cardinality, spectral efficiency and minimal cross-correlation values. Performance analysis of the W/T MPR codes is carried out and found to be superior to other codes. In unipolar 1-D Optical Orthogonal Codes (OOCs) proposed by Salehi et al., the ratio of code length/code weight grows rapidly as the number of users is increased for a reasonable weight. Hence, for a given pulse width, the data rate decreases or in other words for a given data rate very narrow pulses have to be used, because of which dispersion effects will be dominant. To overcome the drawbacks of non-linear effects in large spread sequences of 1-D unipolar codes in FO-CDMA networks, several two-dimensional codes have been proposed. Wavelength-time (W/T) encoding of the two-dimensional codes is practical in FO-CDMA networks. W/T codes reported so far can be classified mainly into two types: 1) hybrid sequences, where one type of sequence is crossed with another to improve the cardinality and correlation properties and 2) matrix codes, 1-D sequences converted to 2-D codes or 2-D codes by construc- tion, to reduce the ’time’ spread of the sequences/codes. Prime-hop and eqc/prime W/T hybrid codes have been proposed where one type of sequence is crossed with another to improve the cardinality and correlation properties. Other constructions deal with conversion of 1-D sequences to 2-D codes either by using Chinese remainder theorem or folding GoLomb rulers. W/T single-pulse-per-row (W/T SPR) codes are 2-D codes constructed using algebraic method Addition Modulo Group operation. Motivation for this work: To design a family of 2-D codes which have the design choice of length of one dimension over the other, and also have better cardinality, spectral efficiency and also low cross-correlation values (thereby have low BER) than that of the reported unipolar 2-D codes. In this thesis, we describe the design principles of W/T MPR codes, for in- coherent FO-CDMA networks, which have good cardinality, spectral efficiency and minimal cross-correlation values. Another feature of the W/T MPR codes is that the aspect ratio can be varied by trade off between wavelength and temporal lengths. We lay down the necessary conditions to be satisfied by W/T MPR codes to have minimal correlation values of unity. We analytically prove the correlation results and also verify by simulation (of the codes) using Matlab software tool. We also discuss the physical implementation of the W/T MPR FO-CDMA network with optical encoding and decoding. We show analytically that when distinct 1-D OOCs of a family are used as the row vectors of a W/T MPR code, it will have off-peak autocorrelation equal to ‘1’. An expression for the upper bound on the cardinality of W/T MPR codes is derived. We also show that 1-D OOCs and W/T SPR codes are the limiting cases of W/T MPR codes. Starting with distinct 1-D OOCs, of a family, as row vectors, we propose a greedy algorithm, for the construction of W/T MPR codes and present the repre- sentations of the results. An entire W/T MPR code family, generated using greedy algorithm, is simulated for various number of interfering users. Performance analysis of the W/T MPR codes and their limiting cases is carried out for various parameter variations such as the dimensions of wavelength, time and weight of the code. We evaluate the performance in terms of BER, capacities of the networks, temporal lengths needed (to achieve a given BER). Multiple access interference (MAI) signal can be reduced, by using a bistable optical hard-limiter device in the W/T MPR code receiver, by eliminating those signal levels which exceed a certain preset level. Performance analysis of the W/T MPR codes and their limiting cases is studied for various parameter variations. For given wavelength × time dimensions, we compare various W/T codes, whose cardinalities are known, and show that W/T MPR family of codes have better cardinality and spectral efficiency than the other (reported) W/T codes. As W/T MPR codes are superior to other W/T codes in terms of cardinality, spectral efficiency, low peak cross-correlation values and at the same time have good performance, makes it a suitable coding scheme for incoherent FO-CDMA access networks.