No description available.

O desempenho energético de edifícios constitui um dos principais fatores de eficiência energética do país. Tal facto é reconhecido pela Comissão Europeia e pelo governo português que legislou o sistema de Certificação Energética dos Edifícios. Neste trabalho, pretende-se estimar o desempenho energético de edifícios residenciais com base em características estruturais e de conjuntura do imóvel, tais como compacidade, áreas de superfície, de parede e de telhado, orientação, área vidrada, carga térmica, etc. O principal objetivo desta dissertação é analisar e caracterizar as circunstâncias que condicionam o desempenho energético, de modo a permitir a identificação de casos que potenciam a diferenciação de estados.O desenvolvimento e este dos algoritmos implementados serão baseados em dados públicos (disponibilizados em UCI Machine Learning Repository / Energy efficiency Data Set), de modo a facilitar a comparação com outros estudos.A ferramenta de classificação a utilizar neste estudo será baseada em sistemas imunológicos artificiais (algoritmos de seleção negativa).

No description available.

No description available.

Supervisory Control and Data Acquisition (SCADA) systems are essential for monitoring and controlling a country's Critical Infrastructures (CI) such as electrical power grids, gas, water supply, and transportation services. These systems used to be mostly isolated and secure, but this is no longer true due to the use of wider and interconnected communication networks to reap benefits such as scalability, reliability, usability, and integration. This architectural change together with the critical importance of these systems made them desirable cyber-attack targets. Just as in other Information Technology (IT) systems, standards and best practices have been developed to provide guidance for SCADA developers to increase the security of their systems against cyber-attacks.With the assistance of EFACEC, this work provides an analysis of a SCADA system under current standards, client requisites, and testing of vulnerabilities in an actual prototype system. Our aim is to provide guidance by example on how to evaluate and improve the security of SCADA systems, using a basic prototype of EFACEC's ScateX# SCADA system, following both a theoretical and practical approach. For the theoretical approach, a list of the most commonly adopted ICS (Industrial Control Systems) and IT standards is compiled, and then sets of a generic client's cybersecurity requisites are analyzed and confronted with the prototype's specifications. A study of the system's architecture is also performed to identify vulnerabilities and non-compliances with both the client's requisites and the standards and, for the identified vulnerabilities, corrective and mitigation measures are suggested. For the practical approach, a threat model was developed to help identify desirable assets on SCADA systems and possible attack vectors that could allow access to such assets. Penetration tests were performed on the prototype in order to validate the attack vectors, to evaluate compliance, and to provide evidence of the effectiveness of the corrective measures.

Facial expression recognition is a particularly interesting field of computer vision since it brings innumerable benefits to our society. Benefits that can be translated into a large number of applications in subjects such as, neuroscience, psychology or computer science. The relevance of the topic is reflected in the vast literature already produced describing notable signs of progress. However, the development and the advancement of new approaches is still facing multiple challenges. Challenges including head-pose variations, illumination variations, identity bias, occlusions, and registration errors. One of the focus in this field is to achieve similar results when moving from a controlled environment to a more naturalistic scenario. Though facial expression recognition has been addressed in considerable different projects, it is feasible to emphasize the call for attention to the design of an interface that simulates addressing patient engagement in healthcare. Since it has been noticed a rising tendency in engaging patients in their healthcare. There are still some open questions need to be answered to make a significant impact on health care.

este protejo integra um sistema de optimização de consumo de energia domésticos de electricidade com integração de um sistema de resposta dinâmica distribuída. começamos por optimizar uma habitação independente, tendo sempre em conta a preferência do utilizador. depois é alargada e aplicada a uma produção distribuída.

No description available.

No description available.

The growing connectivity of vehicles and Vulnerable Road Users, i.e., pedestrians and cyclists, allows to explore solutions based on wireless communication to support safety, efficiency and infotainment applications.However, there are few communication technologies that enjoy similar penetration ratios on cars, bicycles and pedestrians.WiFi is one of such technologies, as can be found in smart phones and in on-board hotspots.This thesis aims to characterize experimentally the wireless link performance and develop a model to estimate the received signal strength (RSS) between WiFi devices installed on bicycles and cars equipped with built-in WiFi APs.The RSS estimation model extends existing empirical models (e.g., the Log-Distance Path Loss model) by including the shadowing of the bicycle-and-cyclist system and of a vehicle.We first characterize the radiation pattern of antennas installed in several mounting points of a bicycle, in order to reduce the set of mounting points to be explored in future measurements.We then measured the radiation pattern of the bicycle and cyclist system, and the radiation pattern of a car with built-in and dedicated WiFi access points.Finally, we evaluate the performance of the model by comparing RSS estimates and measurements collected in selected interaction scenarios between bicycles and car: (i) bicycle overtaking a parked car, (ii) perpendicular crossing with LOS, and (iii) without LOS. We observed that 50% of the RSS estimates our model underestimates by less than are within 10 dBs of measured values about 50% of the RSSI values for the scenarios in LOS, and overestimates the RSSI values by more than 5 DBs about 75% of the RSSI values for the scenario containing obstructions.