FCT: DEE - Teses de Doutoramento
URI permanente para esta coleção:
Navegar
Entradas recentes
- A Distributed Ledger Based Framework for Health Related Data IntegrityPublication . Silva, João Carlos de Fraga Gião da; Sarraipa, João; Gonçalves, Ricardo; Caldas, PauloWith the growing integration of cloud computing and the increasing adoption of Internet-of-Things (IoT) devices, ensuring the integrity and privacy of data has become critical in digital systems. Data integrity is fundamental to maintain the completeness and reliability of data throughout the data lifecycle. Its importance is particularly evident in domains such as healthcare, where accurate diagnoses rely on trustworthy data. However, as systems evolve and become more complex, traditional centralised solutions often lack transparency and resilience, while resource-constrained devices make it more difficult to guarantee security and privacy. This thesis addresses these challenges by proposing a framework that integrates Distributed Ledger Technology (DLT) to support privacy-preserving data sharing and strengthen trust among system stakeholders. Based on this framework, an architecture was designed with three main modules: a middleware integrator for service interoperability, an authorisation manager for fine-grained access control, and a data integrity validator leveraging metadata anchored on a distributed ledger to ensure compliance with the General Data Protection Regulation. A proof of concept was designed and implemented using IoT devices, healthcare data, and low-resource hardware. Experimental results demonstrate that the proposed solution enables efficient data sharing and integrity validation with minimal overhead on the system. The DLT layer validated the integrity of shared data through a metadata model while preserving user privacy. Furthermore, the access control mechanism supported scalable and granular authorisation policies, and the middleware facilitated interoperability across heterogeneous stakeholders. This work contributes to new insights into the security of digital systems and provides responsible entities with a trustworthy approach for sharing data among diverse entities.
- A Design Approach for Collaborative Cyber-Physical SystemsPublication . Nazarenko, Artem Artemovych; Matos, LuísModern Cyber-Physical Systems (CPSs) might contain a large number of components, in fact, being a system-of-systems. These new systems are nowadays characterized by high levels of intelligence and autonomy of their components, requiring collaboration among those compo- nents as well as interaction with outer systems, which led to emergence of the Collaborative Cyber-Physical Systems (CCPSs) concept. CCPSs focus on improving the networking among components and their joint efforts for accomplishing complex tasks, bridging not only the cyber and physical parts, but also the social one. An example of such a joint task is a collabo- rative or complex service delivered either to human users or to other system components. The physical part of a CCPS includes sensing and actuating capabilities deployed in a certain en- vironment and the cyber part provides the abstraction to integrate physical devices and pro- vide high level functionalities, while allowing users to be an active part of CCPS. As for the social aspect, CCPSs in domains such as Smart Homes are expected to be user-centric, with active user involvement from the design phase to the operational phase. Design of the complex CCPS is a non-trivial task that includes both the technical and human aspects. The goal of the designer is to design a system that will address the users’ needs and preferences, while maintaining the agility of the system, so that it can adapt to changing circumstances in a quicker and more reliable way. This is related to the fact that the users’ needs and preferences may change with time, as well as new ones might appear. The goal of this work is to provide the CCPS designer with a design methodology and tools to simplify the design and implementation process. Thus, the designer should be able to build a system skeleton through combining the ready-to-use building blocks in a clear and defined way. This thesis aims to identify and support the qualities and characteristics a design ap- proach needs to effectively aid the CCPS design process, considering collaborative aspects and the growing intelligence of system components. By identifying a gap in high-level methodol- ogies to support the design process of CCPS, this work focuses on design aspects enabling the necessary system-level abstraction. At the same time, mechanisms to ensure the connection between the high-level blueprint and the underlying components, such as specific technolo- gies, protocols, along with other implementation details, are also partially addressed. This ef- fort resulted in the CCPS Design Framework, inspired by the design science research method- ology, which outlines a set of steps to guide the design process. The integration of Digital Twins in the CCPS Design Framework is considered a promising approach to ensure continu- ity between high-level design objectives and their embodiment in real systems. Another key contribution is the extended CCPS-related terminology, which involves rethinking and adapt- ing concepts from the fields of Collaborative Networks, Cyber-Physical Systems, and the In- ternet of Things (IoT) to reflect the specific aspects of CCPS. Moreover, to support the design process, a CPS Service Ontology, extending the capabilities of existing ontologies in the areas of CPS and IoT, is developed, accompanied by a set of tools to assist designers throughout the identified design steps. Finally, the entire approach, including the CCPS Design Framework, design steps, and the proposed tools, is assessed in a Smart Home scenario.
- Accurate Rural Road Network within an Integrated Frame-work of Tools for a Decision Support System in Wildfire ManagementPublication . Lourenço, Miguel Alexandre Gonçalves; Oliveira, Luís; Oliveira, HenriqueThis thesis proposes innovative tools to enhance wildfire management by integrating critical data and addressing limitations in current state-of-the-art systems. Existing Decision Support Systems often use fragmented tools to address different tasks that must be conducted in wildfire scenarios, limiting data integration (such as: rural road networks, several sensor data) and system performance. The research started by developing accurate rural road detection and extraction to overcome current state-of-the-art challenges, notably road occlusions and shadows cast by vegetation or man-made objects, as well as to better interpret roads made of different surface materials. To achieve this, two methods have been utilised over two custom aerial image datasets reflecting real-world scenarios: (i) the first with DeepLabV3+ for road detection, followed by the Zhang-Suen and Guo-Hall thinning algorithms for road extraction; (ii) the second using four U-Net-based architectures that were confronted, incorporating two post-processing techniques to improve rural road detection and extraction. A mobile application has been developed to aggregate data from several hardware sensors, namely atmospheric sensor data for real-time monitoring of firefighting assets and infrared sensor data for post-wildfire hotspot detection, to enhance decision-making in wildfire management. An additional module leverages smartphone sensors and elevation data to geolocate wildfire outbreaks. This application also incorporates a dynamic road network that utilises the proposed road detection and extraction methods and accounts for current road obstructions to optimise route calculations. In addition, a device was developed, integrating a module that provides real-time sensory data during wildfire events and another module that collects sensory data in the wildfire aftermath. Lastly, an interoperable web service has been developed to ensure data interoperability among all developed tools, strengthening situa- tional awareness and leading to well-informed decision-making.
- Fusion of Earth Observation Satellites Data to Improve Forest Height and Aboveground Biomass EstimationPublication . Pires, João Eduardo Albuquerque Martins Pereira; Mora, André; Fonseca, José; Silva, JoãoThe undeniable importance of forests in the sustainability of the planet Earth and its impact on the climate change, increases the interest in monitoring these ecosystems. Within the pa- rameters that support the management of forests there is the Forest Height (FH), which not only provides information by itself, but also can be used as a proxy of other parameters, as the biomass. Remote Sensing observations jointly with machine learning regressors have been commonly used to produce regional and global maps, being calibrated (i.e. fitted) with Light Detection and Ranging (LiDAR) data. In this dissertation a Regression Methodology (RM) was proposed for mapping the FH regionally, with the objective of decreasing the amount of calibration data needed for achiev- ing reliable FH maps. The proposed RM resorts to ALOS-2, Sentinel-1/2 and ancillary data, with robust data preprocessing, feature generation, and features processing processes. The machine learning regressor used in the scope of this dissertation was the extreme gradient boosting which had its results compared to a random forest and a stacking regressor. Then, the same RM, with the addition of the produced FH maps, was used for mapping the above- ground biomass (AGB). This dissertation aims to prove that it is possible to reduce the data needed to calibrate regressors that map the FH and AGB, to demonstrate the advantages of combining different remote sensing datasets, and to show the impacts of including the FH maps in the AGB esti- mation. The main challenge was the development of robust RM processes capable of decreas- ing the amount of calibration data needed to achieve reliable FH maps. This not only decreases the extent of the field or airborne laser scanning campaigns needed, but also it is optimised to be calibrated with spaceborne LiDAR data (namely, from GEDI), which is sparse and limited. The accomplishment of these objectives will allow the frequent production of FH and AGB maps, that are a useful resource for forest management, for wildfire protection, and for the mapping of other important variables.
- Exploring Energy Flexibility in Renewable Energy CommunitiesPublication . Queiroz, Humberto Almeida de; Martins, João; Lopes, RuiTo avoid the worst effects of climate change, there must be a paradigm shift in the way energy is converted, transported, distributed, and consumed. More specifically, there should be a change from the current state, where supply always follows demand, to a new state, where demand is also adjusted instantly to the supply level. In this context, energy flexibility repre- sents an important tool to adapt consumption to the different generation conditions, assuming particular interest in the integration of distributed energy generation systems, based on re- newable sources, in an energy transition towards sustainability. In addition, energy flexibility will be fundamental in solving the challenges imposed by the increasing electrification of so- ciety's processes (e.g., introduction of electric vehicles), which highlights the additional bur- den imposed on energy systems, and the joint effort of society to mitigate changes with the limitation of greenhouse gas emissions. In this context, the contribution of this Doctoral thesis is threefold: firstly, based on a literature review on demand energy flexibility, a novel characterization of the energy flexibil- ity potential of event-based devices is proposed, considering the Portuguese legislation on the operation of renewable energy communities. The second contribution of this study is to pro- pose a new framework to support the management of renewable energy communities, namely the energy sharing process. Lastly, the energy flexibility potential of selected event-based de- vices is explored in a case study to increase the benefits of energy consumers and prosumers and to assess the impact of energy flexibility on energy sharing strategies in renewable energy communities.
- A CONCEPTUAL FRAMEWORK FOR AN EFFECTIVE TECHNOLOGY TRANSFER FROM RESEARCH TO INDUSTRYPublication . Jesus, Elsa Marcelino de; Gonçalves, RicardoNowadays, research plays a crucial role in global economic development. However, a significant challenge exists in effectively transferring research outcomes to the industry. Entrepreneurs often lack knowledge about supporting services and methodologies that can help them evaluate, improve, and validate their ideas or products. This lack of awareness may lead to a natural fear that their ideas, products, or services may not be viable, resulting in many research outcomes remaining unknown or untapped. Therefore, there is a need to create an efficient and effective framework that supports the trans- fer of technology from research to industry, ultimately benefiting the end consumer. In this context, the author contributes to this dissertation with a conceptual framework comprising a set of ideas, studies, guides, and methodologies to assist in the evaluation and validation of implementations and innovations before their launch into a project and the market. It ensures the development of project ideas in alignment with the initial plan, promoting continuous evaluations until a positive result is achieved. This approach facilitates effective exploration and implementation in relation to market needs while maintaining quality in technology transfer. As a result, the proposed framework integrates quantitative and qualitative economic models, learning techniques for project evaluation, and innova- tion mechanisms. This integration guides the creation, development, evaluation, validation, and ex- ploration of research results efficiently and effectively, facilitating the transfer of technology from re- search to industry. The goal is to determine the success of the idea, product, or service in the market in advance, generating income for the industry and benefiting research, industry, and the end consumer. The pro- posed framework components have been successfully demonstrated and validated in various research projects co-financed by the European Commission and the Portuguese government, confirming the validity of all its stages.
- Modelling Smart Manufacturing Assets Targeting Scheduling OptimisationPublication . Alemão, Duarte José Marques; Oliveira, José; Rocha, AndréThe industry sector has evolved faster and faster over the past few decades, driven by increas- ingly complex market demands. Customers now hold greater decision power, and factories are pressured not only to deliver products fast but also to optimize production processes, reducing costs, inefficiencies, and delays. Companies must ensure they can meet customer expectations without compromising operational efficiency. Thus, modern manufacturing systems must be robust and agile, capable of reacting smoothly to external events, and adaptable to unexpected changes. In this world that is be- coming more and more connected, the rise of smart factories, characterized by interconnected and autonomous entities, is transforming how production systems operate. These entities are becoming able to adapt to real-time events but also share critical data between them to opti- mize workflow, minimize downtime, and ensure continuous production. To maintain production efficiency, meet KPIs like makespan, reduce downtimes, and im- prove energy efficiency, or be more prepared for unexpected disturbances in the system, it is important that companies are equipped with robust and adaptable manufacturing scheduling systems. While numerous solutions have been proposed over the years to implement sched- uling systems, in many cases, those approaches focus on specific cases and do not fulfill the necessary requirements to be applied in industry. This research addresses these limitations by providing a more generic, comprehensive, and adaptable approach for smart manufacturing environments. The design and implementa- tion of scheduling solutions in smart manufacturing systems is not standardized and there is not a reference model to develop scheduling solutions that reflect real industrial environments, leading to a gap between reference architectures and scheduling systems. Therefore, the pro- posed research intends to study the main challenges related to manufacturing scheduling and to model manufacturing components targeting the scheduling optimization based on one of the most prosperous reference architectures, RAMI 4.0. Through an extensive literature review, both functional and non-functional requirements were identified and, after analyzing them, the design principles to develop a manufacturing scheduling system were established. Additionally, a methodology was proposed to serve as the foundation for designing scheduling solutions aligned with RAMI4.0, including the identi- fication of the main assets and the development of their corresponding Asset Administration Shells, while addressing key design principles such as data uniformity, KPI harmonization, and automatic rescheduling. Finally, the proposed approach was applied to various use cases, in- cluding the KITT4SME and PERFoRM projects, to demonstrate its efficiency and adaptability. This work aims to fill a critical gap in existing literature but also offers a practical roadmap for industry professionals aiming to fully integrate production scheduling into RAMI4.0, paving the way for smarter, more responsive manufacturing systems in the era of Industry 4.0.
- THE USE OF COOPERATIVE FLEXIBILITY TO IMPROVE THE ENERGY COMMUNITIES’ RESILIENCEPublication . Jesus, Adriana Mar Brazuna de; Martins, João; Pereira, PedroThe increasing integration of renewable energy sources into the power grid has prompted a paradigm shift towards sustainable and resilient energy systems. On the other hand, the energetic flexibility offered by shiftable loads or storage devices brings new win-win solutions for the grid, businesses, households, and the environment. This work explores the concept of Energy Communities (EnCs) cooperative flexibility as a strategic approach to bolstering EnCs resilience. EnCs can influence collaborative efforts among diverse energy stakeholders to optimize energy production, distribution, and consumption. This PhD thesis reviews the key components of EnCs, such as decentralized energy generation, smart grid technologies, and energy flexibility, highlighting their potential to enhance the overall reliability and adaptability of the power grid. The existing literature exhibits a notable gap concerning the EnC resilience. Thus, this research endeavors not only to enhance the resilience of EnCs during faults or power deviations but also to discuss the concept of EnC resilience, incorporating energy flexibility as a pivotal component within the proposed methodology. A community made up of 30 households is considered to conduct a group of use cases, where energy storage system as well as photovoltaic systems are installed. The EnC's resilience is quantified by key metrics, proposed for this thesis, that allow analyzing the community's behavior regarding the user's needs in different situations. The conducted use cases' results show that the proposed Energy Community framework improves the resilience of the community, benefiting not only the community's users as well as the Distribution System Operator (DSO).
- Neuro-Inspired Ultra-low-power CMOS Electronic System (μW range) for ECG and BMI ApplicationsPublication . Teixeira, Miguel de Lima; Goes, João; Príncipe, JoséBrain-machine interfaces (BMIs) require major advances in electronics, so that constrains such as the need of low power, small size, lightweight, and high bandwidth wireless- communications with small data-rates can be solved, Sanchez et al. [1]. Continuous time (CT) asynchronous data converters namely, analog-to-digital converters (ADCs) and analog-to-time converters (ATCs), can be beneficial for certain types of applications, such as, processing of biological signals with sparse information. A particular case of these converters is the integrate-and-fire converter (IFC) that is inspired by the neural system. This dissertation presents and compares two CT asynchronous ATC that do not require an external clock signal. They are two low power IFC solutions, one analog and the other a fully digital dynamic IFC [2, 3]. The first is a closed-loop analog IFC with conventional blocks and on-chip capacitor, although not sacrificing either the chip area or power. The latter, is an open-loop standard cell-based (SCB) IFC, fully synthesizable (with the addition of two on-chip capacitors) and dynamic as each individual block can be powered o↵. Both can be used as an analog frontend (AFE) without requiring external blocks. Being fully-di↵erential - the analog solution is fully-di↵erential, the SCB one is pseudo-di↵erential, it also benefits its performance in AFE applications. As both systems are asynchronous, having a low power dissipation, and with pulse outputs with low data rates, they are a good solution for edge applications, such as low power sensors AFE in internet of things (IoT). Both have been designed and prototyped in a 130 nm CMOS standard process. The analog version has a power dissipation of 53 μW, an energy per pulse of 1060 pJ, and it can convert signals with a peak-to-peak amplitude of 0.6 mV to 2.4 mV and a frequency range of 10 Hz to 4 kHz, and the SCB version: 59 μW, 18 pJ - which is one of the lowest energy per pulse consumption reported for IFC circuits, and 1.6 mV to 32 mV and 2 Hz to 42 kHz, respectively. The maximum pulse density (average firing rate) for analog version is 50 kHz and SCB version 3300 kHz.
- Receiver Design for Reliable and Secure Wireless CommunicationsPublication . Madeira, João Falé de Carvalho; Dinis, Rui; Guerreiro, JoãoWireless networks have become so intrinsically linked with our daily lives to the point it is expected to be always available at any time and place, with high data rates. Due to services such as two factor authentication and remote working, a connectivity problem can severely impact individual productivity, not to mention the potential industrial impacts. To achieve this level of connectivity and speed, many manufacturers compromise on the efficiency and cost of the hardware. In fact, Wi-Fi devices have been known to consume significant power due to the transmission scheme chosen in the protocol, which precludes the usage of saturated power amplifiers. Another challenge lies in the non-ideal nature of the hardware. Manufacturing hardware that operates closer to an ideal transceiver can be significantly more expensive or even impossible. Therefore, there is a large research interest into Digital Signal Processing (DSP) techniques that can provide benefits while still using cheap and efficient hardware. In this thesis the author proposes nonlinear detection techniques that can cope with hardware impairments in the transmitter, or, in the limit, use these effects to achieve a better communication. Since in the current age it is impossible to ignore the security aspects of communication systems, this thesis also presents a study on how these techniques behave in the presence of malicious eavesdroppers.