Dissertação (mestrado) - Universidade Federal de Santa Catarina, Centro Tecnológico, Programa de Pós-Graduação em Engenharia de Automação e Sistemas, Florianópolis, 2016. / Made available in DSpace on 2017-02-14T03:09:15Z (GMT). No. of bitstreams: 1 343672.pdf: 2205846 bytes, checksum: a84be6996ab689a382d1b60ec379c1db (MD5) Previous issue date: 2016 / Dados operacionais são fundamentais para o planejamento e gerenciamento de sistemas de transporte público coletivo. No entanto, o emprego de métodos tradicionais de levantamento de dados apresenta altos custos orçamentários e ineficiência quanto a obtenção de dados como origens e destinos de passageiros. Observa-se, por outro lado, que a crescente popularidade de dispositivos móveis como smartphones possibilita um novo método de levantamento de dados de utilização de ônibus através de uma estratégia de crowdsensing que emprega os dispositivos móveis portados pelos passageiros como plataforma de aquisição de dados.Dentre as tecnologias de comunicação disponibilizadas pelos dispositivos, Bluetooth provê suporte à comunicação de dados com baixo custo e baixo consumo de energia. Neste trabalho, propõe-se um sistema de aquisição de dados operacionais do transporte coletivo através do uso de smartphones portados pelos passageiros. Com o propósito de avaliar a adequabilidade de tal tecnologia no problema de aquisição de dados de ônibus, implementa-se um protótipo, sobre o qual realiza-se uma série de testes de desempenho e consumo de energia. A estratégia de coleta de dados de interesse é avaliada através da utilização do protótipo do aplicativo em simulação de itinerários, onde são obtidos resultados que indicam a viabilidade da arquitetura proposta.<br> / Abstract : Operational data are a key point on planning and maintenance of bus public transit systems. However, the employment of conventional gathering data methods presents high budget costs and inefficiency on gathering data such as origins and destinations of passengers.On the other hand, we note that the growing popularity of mobile devices, such as smartphones, allows a new data gathering method of bus usage through a crowdsensing strategy, which employs the passengers's mobile devices as data gathering plataform.Among the available smartphone communication technologies, Bluetooth provides data communication with low cost and low power.In this work, we propose a data gathering system of public bus usage based on passenger's smartphones. In order to evaluate the proposal's suitability, a prototype is implemented for performance tests and power consumption evaluation. The proposed data acquisition strategy is evaluated through trials of the prototype in itinerary simulations, obtaining results that indicate the viability of our proposal.

abstract: Android has been the dominant platform in which most of the mobile development is being done. By the end of the second quarter of 2014, 84.7 percent of the entire world mobile phones market share had been captured by Android. The Android library internally uses the modified Linux kernel as the part of its stack. The I/O scheduler, is a part of the Linux kernel, responsible for scheduling data requests to the internal and the external memory devices that are attached to the mobile systems. The usage of solid state drives in the Android tablet has also seen a rise owing to its speed of operation and mechanical stability. The I/O schedulers that exist in the present Linux kernel are not better suited for handling solid state drives in particular to exploit the inherent parallelism offered by the solid state drives. The Android provides information to the Linux kernel about the processes running in the foreground and background. Based on this information the kernel decides the process scheduling and the memory management, but no such information exists for the I/O scheduling. Research shows that the resource management could be done better if the operating system is aware of the characteristics of the requester. Thus, there is a need for a better I/O scheduler that could schedule I/O operations based on the application and also exploit the parallelism in the solid state drives. The scheduler proposed through this research does that. It contains two algorithms working in unison one focusing on the solid state drives and the other on the application awareness. The Android application context aware scheduler has the features of increasing the responsiveness of the time sensitive applications and also increases the throughput by parallel scheduling of request in the solid state drive. The suggested scheduler is tested using standard benchmarks and real-time scenarios, the results convey that our scheduler outperforms the existing default completely fair queuing scheduler of the Android. / Dissertation/Thesis / Masters Thesis Computer Science 2014

With technology continuing to advance, a lot of everyday tasks are becoming easier to perform and less physically demanding. However, as a result of this, peoples lives are also becoming more sedentary. There is a need for motivating people to be more active. The project's solution to this problem was to develop a mobile game application which made use of Augmented Reality (AR) technology together with GPS to be able to set the outdoor world as the playing field. Along with using the science of gamification, we strove to create a gameplay that encourages players to continue to play. The resulting game makes the player move outdoors to be able to progress within the game. This feature of the game counteracts sedentary behavior. The results suggest that the use of gamification is an important aspect of making a game that players find fun and motivating. The AR technology was less important and needs more functions than just being able to view 3D-models. / Med fortsatta avancemang inom teknologi, blir vardagliga uppgifter lättare att utföra och mindre fysiskt ansträngande. Dock som ett resultat av detta blir människors vardag mer stillasittande. Det finns ett behov av att motivera människor till att bli mer aktiva. Projektets lösning till detta problem var att utveckla ett spel till mobiltelefoner som använde sig av Augmented Reality teknologi ihop med GPS för att sätta upp världen utomhus som spelplan. Tillsammans med vetenskapen om spelifiering, strävade vi efter att skapa en spelupplevelse som uppmuntrar spelare att fortsätta att spela. Det resulterande spelet kräver att spelare rör på sig utomhus för att kunna nå framsteg inom spelet. Denna aspekt av spelet motverkar stillasittande beteende. Resultaten föreslår att användningen av spelifiering är en viktig aspekt i skapandet av ett spel som spelare finner roligt och motiverande. Användningen av AR-teknologi var mindre viktig och behöver fler funktioner än att bara kunna visa 3D-modeller.

Kor inom mjölkindustrin drabbas ofta av juverinflammationen mastit, som orsakas av en mängd olika bakterier. Eftersom olika bakterier kräver olika behandlingsplaner finns ett behov att identifiera vilken bakterie som har orsakat mastiten. I nuläget finns två sätt att göra detta på, skicka in prover till en veterinär eller utbilda personal på plats för att analysera prover. Båda dessa alternativ är tids- och resurskrävande och kan leda till att mjölkföretagarna väljer att försöka behandla sina kor med antibiotika eller andra onödiga och potentiellt ohållbara metoder. Denna rapport beskriver det arbete som utförts av sju studenter under vårterminen 2018 i samband med kursen TDDD96 - Kandidatprojekt i programvaruutveckling. I projektet utvecklades en prototyp för ett system som med hjälp av ett neuralt nätverk klassificerar bakterier hos mastitdrabbade kor utifrån bilder på bakterieodlingar. Som gränssnitt mellan mjölkföretagare och klassificerare utvecklades en Android-applikation där slutanvändaren lätt kan ta bilder och mata in relevant information för att sedan skicka dessa till en server. Under implementationen av produkten delades projektgruppen in i tre delgrupper, varje delgrupp arbetade på ett av de tre delsystemen applikation, server och klassificerare. Projektgruppen använde sig av en Scrum-liknande arbetsmetodik där utvecklingsfasen delades upp i olika sprints. Den klassificerare som tagits fram har lyckats få en klassifikationssäkerhet som är i underkant med veterinärerna. Dock har klassificeraren stor potential och kan utvecklas till att vara en bra lösning till det presenterade problemet inom mjölkindustrin.

Orientador: Aparecido Nilceu Marana / Banca: João Eduardo Machado Perea / Banca: Fátima de Lourdes dos Santos Nunes Marques / Resumo: Na era da informação em que vivemos a demanda por conectividade e acesso a dados é crescente. Neste cenário os dispositivos móveis estão se tornando cada vez mais populares. Com o aumento da demanda por este tipo de dispositivo aumenta também o volume de informações neles armazenadas. Com isso, os dispositivos móveis estão se tornando uma importante fonte de informação que precisam de mecanismos adequados de proteção. O objetivo desta dissertação de mestrado foi promover um estudo sobre a viabilidade da utilização de Biometria para a proteção de informações dos usuários em dispositivos móveis, em particular nos smartphones, via reconhecimento de faces. A opção pelo reconhecimento facial deveu-se ao fato dos smartphones proverem naturalmente recursos de hardware para a aquisição de vídeos. Para a realização do estudo proposto, foi desenvolvido um sistema de reconhecimento facial na plataforma Android, denominado BioMobile, cuja arquitetura foi projetada para permitir a execução integralmente no domínio dos dispositivos móveis. Para a detecção das faces, foi adotado o algoritmo Viola-Jones, enquanto que para o reconhecimento foram usados os descritores PCA e LBP. O BioMobile foi desenvolvido para operar nos modos de autenticação e de reconhecimento. Além disso, ele trabalha com vídeos, o que possibilita a adoção da técnica de maioria de votos, tornando-o mais tolerante a falhas e mais preciso. Os resultados experimentais mostraram que, ao contrário do algoritmo baseado em LBP, o algoritmo baseado em PCA se torna inviável quando usado em dispositivos com restrições de memória e de processamento. Os resultados obtidos também indicam que o modo de operação de reconhecimento quando aplicado totalmente no domínio dos dispositivos móveis não é conveniente em termos... (Resumo completo, clicar acesso eletrônico abaixo) / Abstract: In the information age in which we live the demand for connectivity and data access is growing. In this scenario the mobile devices are becoming increasingly popular, especially the smart phones. The increasing demand for this type of device leads to high personal and professional information storage. Therefore, mobile devices are becoming important sources of information that need proper protection mechanisms. The purpose of this master dissertation was to carry out a study on the feasibility of using biometrics as a means of safe, efficient and appropriate information protection on mobile devices, especially in smart phones, via user identification through face recognition. The choice for facial recognition was due to the fact that the smart phones provide embedded video cameras. In order to conduct the proposed study, it was developed a face recognition system on the Android platform, called BioMobile, whose architecture was designed to allow its full execution without the need of external servers. For face detection it was adopted the Viola-Jones algorithm, while for face recognition it were assessed the PCA and LBP descriptors. BioMobile was also designed to operate in authentication and recognition modes. Besides, it works with videos instead of still images that allow the adoption of the technique of majority voting becoming it more fault-tolerant and precise. The experimental results showed that the PCA algorithm is not feasible when used in devices with memory and processing power limitations. LBP, in contrast, showed to be appropriated. The results also indicated that the recognition operation mode fully executed in the mobile devices is not so convenient due to the processing time required. On the other hand, the authentication operation mode showed good performance regarding this issue. The system also showed few error rates (around 2%) depending on the adopted configuration / Mestre

Os aplicativos móveis, desenvolvidos originalmente para a área do entretenimento, hoje estão presentes nos mais diversos domínios, sendo comuns inclusive em áreas de alto valor agregado, como: varejista, logística, bancária, médica, entre outras. Portanto, a qualidade e correção dos aplicativos móveis tornam-se obrigatórios e as atividades de teste essenciais. Porém a qualidade das aplicações móveis nem sempre é satisfatória. Isso ocorre devido ao fato dessas aplicações sofrerem com a pressão do mercado e passarem por um processo muito rápido de desenvolvimento, onde geralmente a fase de testes é negligenciada ou realizada de forma superficial, pela própria equipe de desenvolvimento, comprometendo assim a qualidade da aplicação. Este trabalho propõe uma abordagem baseada no Behavior Driven Development para ajudar na definição de testes de sistema para aplicativos nativos do Android. A abordagem proposta utiliza os arquivos de leiaute da aplicação para extrair informações sobre os componentes da interface e sobre os eventos esperados pelo sistema. A partir dessas informações, é possível verificar a cobertura dos cenários existentes em relação aos eventos disponíveis na interface com o usuário. Além disso, é possível identificar elementos do leiaute que não são exercitados pelos cenários existentes. A abordagem proposta é implementada por uma ferramenta chamada Android Behavior Testing Tool que, por meio da interpretação dos cenários do Behavior Driven Development, fornece uma visão geral do fluxo comportamental da aplicação ao testador (visão hoje não disponível), proporcionando assim uma noção de fácil compreensão sobre a cobertura dos testes em relação aos elementos da interface do aplicativo. Desta forma, o testador pode julgar a integridade dos casos de teste disponíveis em relação às funcionalidades implementadas e, se necessário, implementar novos testes. A ferramenta também faz uso dos arquivos de leiaute do aplicativo para identificar os componentes da interface que não foram testados e gera, neste caso, modelos de cenários no formato do BDD, automatizando assim a tarefa de escrita dos mesmos. A abordagem proposta foi utilizada em quatro aplicativos Android e se mostrou útil, uma vez que, em três estudos de caso foram detectados bugs oriundos de inconsistências lógicas nos cenários ou elementos não exercitados pelos cenários. / Mobile applications, originally developed for entertainment, nowadays are present in a wide range of domains, being common even in areas of high value such as retailer, logistics, banking, and medical, among others. However, the quality and correctness of mobile applications become mandatory and testing activities are essential. However, the quality of mobile applications is not always good enough. This is because these applications suffer from market pressure and pass through a very rapid development process where the testing phase usually is neglected or superficially performed by the development team itself, thus compromising the quality of the application. This work proposes an approach based on Behavior Driven Development to help to define system tests for native Android applications. The proposed approach uses the application's layout files to extract information about the interface components and the events expected by the system. From this information, it is possible to check out the coverage of existing test scenarios against events available in the user interface. In addition, it is possible to identify unexercised usage scenarios from the existing test scenarios. The proposed approach is implemented by a tool called Android Behavior Testing Tool which, through the interpretation of the BDD usage scenarios, provides to the tester an overview of the behavioral flow of the application (otherwise unavailable), thus providing a notion of easy understanding of test coverage in relation to the application interface elements. In this way, the tester can judge the integrity of the available test cases in relation to the functionalities implemented and, if necessary, implement new tests. The tool also makes use of the application's layout files to identify untested interface components and in this case generates test scenario models in the BDD format, thus automating the writing task of the scenarios. The proposed approach was used in four Android applications and proved to be useful, since in three case studies bugs were detected. Detected bugs originated from logical inconsistencies in the test scenarios or elements that were not exercised by the scenarios.

abstract: Debugging is a boring, tedious, time consuming but inevitable step of software development and debugging multiple threaded applications with user interactions is even more complicated. Since concurrency and synchronism are normal features in Android mobile applications, the order of thread execution may vary in every run even with the same input. To make things worse, the target erroneous cases may happen just in a few specific runs. Besides, the randomness of user interactions makes the whole debugging procedure more unpredictable. Thus, debugging a multiple threaded application is a tough and challenging task. This thesis introduces a replay mechanism for debugging user interactive multiple threaded Android applications. The approach is based on the 'Lamport Clock' concept, 'Event Driven' implementation and 'Client-Server' architecture. The debugger tool described in this thesis provides a user controlled debugging environment where users or developers are allowed to use modified record application to generate a log file. During the record time, all the necessary events like thread creation, synchronization and user input are recorded. Therefore, based on the information contained in the generated log files, the debugger tool can replay the application off-line since log files provide the deterministic order of execution. In this case, user or developers can replay an application as many times as they need to pinpoint the errors in the applications. / Dissertation/Thesis / M.S. Computer Science 2012

The project described in this thesis implements a position based offer system that allows the user to receive offers when they enter zonesspecified in a database. The goals associated with this project were: to find and evaluate products with similar functionalities and to design and implement a working prototype. The system consists of four main parts, a smartphone, a server, a website, and a billing system. The smartphone is based upon the Android operating system and uses the global positioning system (GPS) to track its position. The server is the brain of the system and all the other devices communicate through it.The server runs JBoss and has a connection to an Oracle database through hibernate. JavaServer Pages (JSP) is used by the server to generate the website and to connect the Java methods with the HTML-forms at the website. User privacy is one aspect that has been investigated in relation to this project. The user should have options to choose the information that is to be shared and stored. When storing private information it is important to secure the information, preferably by encryption and a log in system. A small program was developed to test the scalability of the server and the result showed that the server could cope with 1000 simultaneous users polling the server with a 5 second interval.This will, however, increase several times when a dedicated server machine is used. Today there is no other product at the market that has a connection to the billing system; this means that this product will have a huge advantage and market potential compared to similar products.

Social media is becoming a bigger part of society and has now become a regular part of the worldof professional companies. Incentive is a social intranet for companies, developed by Mindroute inSweden. Incentive is available online but for the moment not as an application for Android. This thesis focuses on developing a prototype Android application for Incentive, evaluating theusability of the result. Two designs were developed and the study showed that while both weregood and user friendly a combination of the two should be considered.

Augmented Reality (AR) can be described as one of possible steps between real world and fully virtual reality. Into this mixed reality we can make an overlay with virtual objects onto the real world typically by capturing camera images in real-time to produce a new layer to the environment with which we can interact. Mobile Augmented Reality (MAR) is a term used when equipment through which we achieve AR is small in size and typically easy to carry e.g. a smartphone or a tablet. The concept of using AR in facilitating learning and improving its quality seems to attract more attention in the academic world in recent years. One of the areas that receive much attention is AR language learning. In this thesis an experiment on a group of 20 people was conducted to answer the question: “Is MAR language learning system a viable solution for language learning?” For the purpose of the experiment an AR Language Learning Tool was designed for Android smartphones. This AR Language Learning Tool facilitated vocabulary learning by displaying 3D objects along with their spelling and providing audio of pronunciation. Participants were divided into an equal control group and test group. The control group learned new vocabulary through classic flashcards while the test group used the previously designed AR Language Learning Tool. The Vocabulary Knowledge Scale questionnaires were provided for both groups right after learning and one week later. By performing statistical analysis with Student’s t-test on gathered data it was discovered that there is a positive improvement in long term recall rate in the AR Language Learning Tool group when compared with the flashcards learning group. No difference was found in short term recall rate between both groups. Participants also provided feedback about their quality of experience and enthusiasm for new learning methods. Their answers were very positive and provided proof that mobile AR is a viable method of learning vocabulary.