Browsing by Issue Date, starting with "2015-10-15"
Now showing 1 - 2 of 2
Results Per Page
Sort Options
- A Platform to Support the Development of Applications Based on the Segmentation and Labelling AlgorithmPublication . Guia, Sabrina Santos; Espírito Santo, António Eduardo Vitória do; Paciello, VincenzoThe need for networks of sensor that operate in an increasingly efficient way is contributing to boost the research methods with the aiming of find the best methods about how to manage the information. At the same time, the technological explosion that is taking care of our daily life, have influence not only on what is likely to be standardized, but also in the possibility of execution. An example of this is the need to exchange the knowledge about how the sensor’s signal behaves in a standardized manner, as well as, some of its features and isolated parameters. The compaction of data to be managed and transmitted clearly becomes useful, especially when there is interest in the study of signal characteristics, such as; amplitude; the presence of noise; steady state values tendency, etc. Small parameter capable of induce the general behaviour of the measured signal. In the world of sensors, intelligence is focused on the same point of measurement made by the transducer. From this point, three algorithms are presented. The first is based on the rapid measurement of procedural techniques of the Fourier transform (FFT), while the second is based on the compressive sampling theory (CS). Finally it is presented a proposed algorithm, in the time domain, based on the processes of segmentation and labelling (S&L) of a sampled signal that is also proposed for defining the new IEEE 1451 standard. The main goal of this work is focused on the management of a smaller amount of data, but the acquisition of the same amount and quality of information. Thus, the ideal is that the signal is sampled in order to create redundancy between neighbouring samples. It is therefore not necessary segmentation of each pair of samples, storing and transferring only the samples that are held to bring important information. As a result of the segmentation process, two vectors are obtained for storing amplitudes and time indices of samples and labelling process, the segments are classified according to their behaviour in eight different classes and stored in a third vector. This set of MCT vectors offers a structurally standardized platform that supports sensor’s data exchange. The reconstruction of the acquired signal is also allowed by this structure. The method was implemented and each function and their respective operating modes are described in detail. In addition, input and output parameters of each function are also described. Afterwards, the project was prepared to implement in a microcontroller, whose architecture from ARM. In order to demonstrate the performance of the proposed algorithm, experimental results in the area of instrumentation, analysis and signal processing, are reported and exposed. After the execution of the above-mentioned procedure, the sequence segments analysis reveals that the algorithm has the ability to extract global information form the acquired signals, such as a human observer. In addition, its low computational cost allows the inclusion of the proposed method in smart sensors in a variety of application, enabling execution in real time. The document is divided into four chapters. In the first chapter is made a brief research concerning the state of the art related with the main subject of this work. Theoretical considerations concerning knowledge extraction from sample, analysis and data processing are presented. The best known algorithms are described in the second chapter special focus is given observing the application area. In the third chapter software and hardware tools used in the implementation process are described. The fourth chapter carries out the implementation of the proposed algorithm and described the respective implementation. The members of API functions are individually tested and the results presented and analysed. In the fifth, and final, chapter final observations are drawn with conclusions. Suggestion about future work is also presented.
- Selfies in a lonely crowd: interaction and meaning in social networksPublication . Correia, João CarlosIn the early days of the Internet, significant researchers on sociological aspects relating to the Internet thought that online and off life worlds were somewhat disconnected as if they were parallel universes. Cyberculture emphasized a more contingent, flexible and fluid experience of selfhood. Those approaches produced a strong influence in early postmodern Internet studies (Hayward, 1991; Stone, 1996; Mosco, 2004; Talbot, 1995). Reading those earlier papers, it seems that the Internet has opened the way for the full realization of political utopias that conceived cyberspace as having an inherently democratic structure (Barlow, 1996). The web of our days is far from being the same of earlier stages. Dominated not just by Google, Twitter, YouTube, and Wikipedia, it has recently been shaped by WikiLeaks and the so-called Facebook and Twitter revolutions in North Africa and the Middle East. Over the years, perspectives on identities and social interactions have very much changed. Today, Internet studies follow a path based on the empirical research of the effective social interactions performed in the digital environment. Most of these perspectives derive its theoretical framework from important currents of social thought such as phenomenological sociology and symbolic interactionism. The purpose of this paper is to discuss complex literature on the issue of the autonomy of the self with relevance to understanding the meaning of political activism and mobilization.