This thesis deals with the design and implementation of autonomous robot with using of the platform ROS. Its goal is to get to know the ROS and use it to implement autonomous control of real robot Leela.

This thesis deals with realization of modular camera system. Introductory part is focused on market research and concept of the system. Next part is dedicated to the choice of components for realization of the system. Practical part deals firstly with design of printed circuit board for supporting control unit which will be controlling peripherals like motion sensor, IR illuminator, eventually magnetic door contact and also focuses on software creation for main and supporting control unit.

The master’s thesis deals with the design of the Smart Home System (SHS), which takes advantage of the 'Raspberry Pi' single-board computer. Background research about the theoretical concept of SHS is carried out, which reveals weaknesses in that field. The aim of the thesis is elimination these weak points and takes into account reliability, extensibility and low acquisition price. The practical part is introduced by design of particular modules, which include both hardware design and software. The project is concluded with integration all components into the single functional universal system together with the extensibility presentation.

This thesis project aims for the development of a demonstration platform on mobile devices for testing and demonstrating algorithms for 2D and 3D spatial sound reproduction. The demo system consists of four omnidirectional microphones in a square planar array, an Octo sound card (from Audio Injector), a Raspberry Pi 3B+ (R-Pi) single-board computer and an inertial measurement unit (IMU) located in the center of the array. The microphone array captures sound, which is then digitized, and in turn, transferred to the R-Pi. On the R-Pi, the digitized sound signal is rendered through the directional audio coding (DirAC) algorithm to maintain the spatial properties of the sound. Finally, the digital signal and spatial properties are rendered through Dirac VR to maintain a spatial stereo signal of the recorded environment. The directional audio coding algorithm was initially implemented in Matlab and then ported to C++ since the R-Pi does not support Matlab natively. The ported algorithm was verified on a four-channel in and six-channel out system, processing 400 000 samples at 44 100 kHz. The results show that the C++ DirAC implementation maintained a maximum error of 4.43e-05 or -87 dB compares to the original Matlab implementation. For future research on spatial audio reproduction, a four-microphone smartphone mock-up was constructed based on the same hardware used in the demo system. A software interface was also implemented for transferring the microphone recordings and the orientation of the mock-up to Matlab.

Orientador: Carlos Eduardo Angeli Furlani / Resumo: A agricultura digital tem contribuído com a melhoria da eficiência na aplicação de insumos ou no plantio em local pré-determinado, resultando no aumento da produtividade. Nesta realidade a aplicação de técnicas de Processamento de Imagens Digitais, bem como a utilização de sistemas que utilizam a Inteligência Artificial, tem ganhado cada vez mais a atenção de pesquisadores que buscam a sua aplicação nos mais diversos meios. Com o objetivo de desenvolver um sistema robótico que utiliza um sistema de visão computacional capaz analisar uma imagem e, detectar basicamente a presença de cana-de-açúcar e planta daninha, bem como a ausência de qualquer planta, o projeto desenvolvido unificou conhecimentos sobre estas duas áreas da ciência da computação com a área de robótica e agricultura que, culminou no desenvolvimento de uma estrutura robótica com ferramentas gratuitas, como é o caso dos softwares e hardwares modulares voltados para o ensino de informática em escolas. A união de tudo isso resultou em uma estrutura de software e hardware que captura e armazena imagens em um banco de dados; além de possibilitar a classificação de imagens pelos usuários habilitados por meio de aplicativo Android. Por meio da verificação da acurácia entregue pelos algoritmos de Machine Learning, com injeção cíclica e, pela análise do tempo de resposta, foi constatado que o sistema é capaz, munido destas informações, de gerar classificadores que, remotamente são carregados pelo DRR (Dispositivo Robótic... (Resumo completo, clicar acesso eletrônico abaixo) / Abstract: Digital agriculture has contributed to improving efficiency in the application of inputs or planting in a predetermined location, resulting in increased productivity. In this reality, the application of Digital Image Processing techniques, as well as the use of systems that use Artificial Intelligence, has increasingly gained the attention of researchers who seek their application in the most diverse media. In order to develop a robotic system capable of creating a computer vision system capable of analyzing an image and basically detecting the presence of sugarcane and weed, as well as the absence of any plant, the project developed unified knowledge on these two areas of computer science with the area of robotics and agriculture, which culminated in the development of a robotic structure with free tools, such as software and modular hardware aimed at teaching computer science in schools. The combination of all this resulted in a software and hardware structure capable of allowing the capture and storage of images in a database; in addition to enabling the classification of images by users enabled through the Android application. By checking the accuracy delivered by the Machine Learning algorithms with cyclic injection and analyzing the response time, it was found that the system was able, with this information, to generate classifiers that are remotely loaded by the RRD and these, in turn, were able to classify images in sugarcane fields in real time. / Mestre

This project revolves around creating a home automation system prototype with the main focus being the ability to lock/unlock a door through the internet. The system consists of a central device, a server and an Android application.The central device is a microprocessor, in this case, a Raspberry Pi that connects to the Internet and receives an order to control a motor which in turn turns the lock with the help of gears. The ability to rotate the motor in both directions is achieved by the use of an H-bridge. The server manages users and devices, and handles the communication between the application and the central device. Users and devices are stored in a database on the server. The application is a frontend which presents the user with a list of devices to interact with.The main prototype where the Raspberry Pi acted as a central device was abandoned due to time and resource constraints. It was instead used to control the motor directly. This brought up some problems concerning powering the device using batteries. The software of the prototype is mostly working but due to the same time limitations not all planned features could be implemented.

Machine-to-Machine (M2M) communications are used for several purposes, forinstance to transmit information derived from measurements collected frommonitoring instruments. M2M communications also allow intelligent devices toexchange real-time data without human intervention. Through a literaturesurvey regarding M2M, Mobile Communications, and Communication Protocolsfor M2M, such as the Constrained Application Protocol (CoAP), we found thatthe CoAP-UDP model is more suitable for M2M systems, than the HTTP-TCPapproach. Additionally, CoAP supports a DTLS implementation to provide endto-end security to protect communications. Consequently, CoAP was the selectedtechnology that allowed us to achieve the goal of designing a low-cost, scalable,secure, and standard-based communication solution for the company supportingthe project: Solelia Greentech. This company is the largest provider inScandinavia of solar chargers for electrical vehicles. The development andexperimental implementation of this solution was also successfully accomplished.We created a prototype that is able to gather information from a pulse generator(e.g. smart meter), process the data, run a CoAP server, and transmit dataresources to CoAP clients through a secure DTLS channel. Furthermore, aperformance analysis of the system and other existing Web server alternativeswas performed. As a result of this process, we concluded that the CoAP serverwe developed reaches between four and seven times higher throughputs than thecompared systems. Therefore, this project represents a viable alternative forexisting solutions on the market. / Machine-to-machine (M2M) kommunikation används för flera syften, till exempel överföra information från mätningar som samlats in från övervakningsprogram instrument. M2M kommunikation gör det också möjligt att intelligenta enheter utbyter data i realtid utan mänsklig inblandning. Genom en litteraturstudie om M2M, mobil kommunikation, och kommunikationsprotokoll för M2M, såsom Constrained Application Protocol (CoAP), fann vi att CoAP-UDP-modellen är mer lämpade för M2M-system, än HTTP-TCP strategi. Dessutom, CoAP stöder ett DTLS genomförande som bidrar med end-to-end säkerhet för att skydda kommunikation. Följaktligen CoAP var den valda tekniken som tillät oss att uppnå målet att utforma en billig, skalbar, säker och standardbaserad kommunikationslösning för företag som stödde projektet: Solelia Greentech. Detta företag är den största leverantören i Skandinavien av solar laddare för eldrivna fordon. Utveckling och experimentella genomförande av denna lösning var också lyckat fulländad. Vi skapade en prototyp som kan samla information från en pulsgenerator (t.ex. smarta mätare), process data, köra en CoAP server, och överföra dataresurser till CoAP-klient genom en säker DTLS kanal. En prestandaanalys av systemet och andra befintliga webbservern alternativ utfördes. Som en följd av denna process, vi drog slutsatsen att CoAP servern vi utvecklat når mellan fyra och sju gånger högre genomloppstid än de jämförda systemen. Därför Detta projekt är ett lönsamt alternativ för befintliga lösningar på marknaden.

The development of innovative applications of Artificial Intelligence (AI) is inseparable from the sharing of public data. However, as people strengthen their awareness of the protection of personal data privacy, it is more and more difficult to collect data from multiple data sources and there is also a risk of leakage in unified data management. But neural networks need a lot of data for model learning and analysis. Federated learning (FL) can solve the above difficulties. It allows the server to learn from the local data of multiple clients without collecting them. This thesis mainly deploys FL on the Raspberry Pi (RPi) and achieves federated averaging (FedAvg) as aggregation method. First in the simulation, we compare the difference between FL and centralized learning (CL). Then we build a reliable communication system based on socket on testbed and implement FL on those devices. In addition, the Paillier encryption algorithm is configured for the communication in FL to avoid model parameters being exposed to public network directly. In other words, the project builds a complete and secure FL system based on hardware. / Utvecklingen av innovativa applikationer för artificiell intelligens (AI) är oskiljaktig från delning av offentlig data. Men eftersom människor stärker sin medvetenhet om skyddet av personuppgiftsskydd är det allt svårare att samla in data från flera datakällor och det finns också risk för läckage i enhetlig datahantering. Men neurala nätverk behöver mycket data för modellinlärning och analys. Federated learning (FL) kan lösa ovanstående svårigheter. Det gör det möjligt för servern att lära av lokala klientdata utan att samla in dem. Denna avhandling använder huvudsakligen FL på Raspberry Pi (RPi) och uppnår federerad genomsnitt (FedAvg) som aggregeringsmetod. Först i simuleringen jämför vi skillnaden mellan FL och CL. Sedan bygger vi ett pålitligt kommunikationssystem baserat på uttag på testbädd och implementerar FL på dessa enheter. Dessutom är Paillier -krypteringsalgoritmen konfigurerad för kommunikation i FL för att undvika att modellparametrar exponeras för det offentliga nätverket direkt. Med andra ord bygger projektet ett komplett och säkert FL -system baserat på hårdvara.

Robotics will be a dominant area in society throughout future generations. Nowadays its presence is increasing in the majority of contexts of daily life, with devices and mechanisms which facilitate the accomplishment of diverse daily tasks; as well as at labor level, where machines occupy more and more jobs. This increase in the presence of autonomous robotic systems in society is due to the great efficiency and security they offer compared to human capacity, thanks mainly to the enormous precision of their sensor and actuator systems. Among these, vision sensors are of utmost importance. Humans and many animals enjoy powerful perception systems in a natural way, but which in Robotics constitutes a constant line of research. The main problem lies in the correct interpretation of visual data and the extraction of relevant information from camera images. Thus, Robotics becomes something beyond an scientific are, but also a social and cultural topic. Therefore, it is essential to raise an early awareness and train younger students to acquire the skills which will be most demanded in the short and mid-term future. In doing so, we will be ensuring their integration into a labor market dominated by intelligent robotic systems. In addition to having a high capacity for reasoning and decision-making, these robots incorporate important advances in their perceptual systems, allowing them to interact effectively in the working environments of this new industrial revolution. Since a few years ago, there are different Educational Robotics kits available in the market which are designed to be used in pre-university education. To use them as a learning tool, a correct teacher training is necessary, as well as a change in the teaching-learning methodology and in the educational environment in general. In addition, taking into account that young people live immersed in a constant environment of technological learning, most of these kits usually have a short period of interest for students, who demand motivating intellectual challenges. This thesis aims to provide several solutions to some classic problems inherent to Robotics, such as navigation and localization, but using a camera as the main sensor. In addition, a learning framework for teaching of Robotics with Vision as a subject is presented. Using it the students at pre-university curricular level learn the principles of Science and Engineering and the computer programming skills demanded in today's society. The use of Python language and its exercises about robots with vision makes this learning framework unique and more powerful than other existing frameworks. This teaching framework has been successfully used in several secondary education schools during the last two academic years (2016/2017 and 2017/2018), which includes: its software infrastructure, its hardware platform, an academic curriculum with theoretical and practical content, as well as a constructivist pedagogical methodology. The performance and satisfaction of more than 2,000 students and teachers using it, in curricular subjects such as Programming, Robotics and Technology and ICTs of Secondary Education (CSO) and extracurricular activities, have been evaluated.

Ansiktsigenkänning med hjälp av maskininlärning är ett växande område och används i många sammanhang i dagens samhälle, till exempel som autentiseringsmetod i mobiltelefoner. De flesta system för ansiktsigenkänning har haft stor budget och starka utvecklare bakom sig, men går det att skapa ett fungerande system med begränsade resurser och datamängd? Det här projektet undersöker hur mycket data som krävs för att producera en fungerande ansiktsautentiseringssmodul för kontorsmiljö baserad på bildklassificering. I projektet används ett förtränat Convolutional Neural Network (ResNet34), data som är insamlad med hjälp av uppdragsgivaren samt en bilddatabas från NVIDIA. Resultaten visar att mängden data som krävs för att producera en tillförlitlig modell troligtvis överstiger den mängd som är rimlig att samla in från användaren. / Face recognition using machine learning is a changing field and is used in many contexts in today’s society, for example as an authentication method in mobile phones. Most face recognition systems have had large budgets and strong developers behind them, but is it possible to create a working system with a limited amount of resourses and data? This project investigates how much data is required to produce a working face recognition module for an office environment based on image classification. This project used a pretrained Convolutional Neural Network (ResNet34), data collected with the help of the client, and an image database from NVIDIA. The results show that the amount of data required to produce and reliable model probably exceeds the amount that is reasonable to collect from the user.