Background.As a result of the digital era, vehicles are being digitalised in arapid pace. Autonomous vehicles and powerful infotainment systems are justparts of what is evolving within the vehicles. These systems require more in-formation to be transferred within the vehicle networks. As a solution for this,Ethernet was suggested. However, Ethernet is a ’best effort’ protocol which cannot be considered reliable. To solve that issue, specific implementations weredone to create Automotive Ethernet. However, the out-of-the-box vulnerabil-ities from Ethernet persist and need to be mitigated in a way that is suitableto the automotive domain. Objectives.This thesis investigates the vulnerabilities of Ethernet out-of-the-box and identify which vulnerabilities cause the largest threat in regard tothe safety of human lives and property. When such vulnerabilities are iden-tified, possible mitigation methods using security measures are investigated.Once two security measures are selected, an experiment is conducted to see ifthose can manage the latency requirements. Methods.To achieve the goals of this thesis, literature studies were conductedto learn of any vulnerabilities and possible mitigation. Then, those results areused in an OMNeT++experiment making it possible to record latency in a sim-ple automotive topology and then add the selected security measures to get atotal latency. This latency must be less than 10 ms to be considered safe in cars. Results. In the simulation, the baseline communication is found to take1.14957±0.02053 ms. When adding a security measure latency, the total dura-tion is found. For Hash-based Message Authentication Code (HMAC)-SecureHash Algorithm (SHA)-512 the total duration is 1.192274 ms using the up-per confidence interval. Elliptic Curve Digital Signature Algorithm (ECDSA)- ED25519 has the total latency of 3.108424 ms using the upper confidenceinterval. Conclusions. According to the results, both HMAC-SHA-512 and ECDSA- ED25519 are valid choices to implement as a integrity and authenticity secu-rity measure. However, these results are based on a simulation and should beverified using physical hardware to ensure that these measures are valid. / Bakgrund.Som en påföljd av den digitala eran, så har fordon blivit digitalis-erade i ett hastigt tempo. Självkörande bilar och kraftfulla infotainmentsys-tem är bara några få av förändringarna som sker med bilarna. Dessa systemkräver att mer information skickas genom fordonets nätverk. För att nå dessahastigheter föreslogs Ethernet. Dock så är Ethernet ett så kallat ’best-effort’protokoll, vilket inte kan garantera tillförlitlig leverans av meddelanden. För attlösa detta har speciella tillämpningar skett, vilket skapar Automotive Ethernet.Det finns fortfarande sårbarheterna av Ethernet kvar, och behöver hanteras föratt tillämpningen skall vara lämplig för fordonsindustrin. Syfte.Denna studie undersöker vilka sårbarheter som finns i Ethernet ’out-of-the-box’ och identifierar vilka sårbarheter som har värst konsekvenser urperspektivet säkerhet för människor och egendom. Två säkerhetsimplementa-tioner väljs ut för att se över vidare de kan användas för kommunikation i bilar. Metod.För att nå arbetets mål, så genomfördes en literaturstudie för attundersöka sårbarheter och potentiella motverkningar. Studiens resulat använ-des sedan i en simulering för att kunna mäta fördröjningen av en enkel topologii en OMNeT++miljö. Sedan addera den tiden med exekveringstiden för säker-hetsimplementationerna för att få en total fördröjning. Kommunikationstidenmåste vara mindre än 10 ms för att räknas som säker för bilar. Resultat.I simuleringen, så ger mätningarna en basal kommunikation på1.14957±0.02053 ms. När säkerhetsimplementationerna tillsätts så får manden totala kommunikationstiden. För HMAC-SHA-512 mäts den totala kom-munikationstiden till 1.192274 ms genom att använda den övre gränsen av kon-fidensintervallet. För ECDSA - ED25519 mäts tiden till 3.108424 ms. Slutsatser.Enligt resultaten så är både HMAC-SHA-512 och ECDSA - ED25519möjliga alternativ för integritets- och äkthetstillämpningar i fordorns kommu-nikation. Dessa resultaten är dock framtagna ur en simulering och bör verifierasmed hjälp av fysisk hårdvara så mätningarna är sanningsenliga.

Niche partitioning is a major component in understanding community ecology and how different species divide limited environmental resources, enabling them to coexist. Temporal niche partitioning has been widely studied in a broad sense, such as in species that forage on similar nutritional sources dividing activity along diurnal and nocturnal classifications. Here, we approach this temporal niche partitioning with higher resolution to investigate partitioning between species within the same broad temporal and foraging niche. Two species of diurnal orb-weaving spiders (Araneae: Araneidae), Verrucosa arenata and Micrathena gracilis, both construct their orbs in spatially similar locations throughout the understory of deciduous forests in the morning, forage on flying insects throughout the day, and retreat in the evening. However, despite consisting of what appear to be roughly similar total lengths of adhesive silk in the capture spiral, overall orb structure is starkly different: V. arenata orbs are relatively large in diameter and sparse with capture threads; M. gracilis orbs, condensed in diameter and tightly coiled. What other differences might distinguish foraging strategy within this same niche? With extensive observation in their natural environment, we have found that these two species employ two distinct strategies by modulating behavior and orb structure: V. arenata construct orbs earlier in the day, resulting in a longer foraging period. However, V. arenata webs are more likely to be destroyed during the day such that there is a higher variance in foraging duration in V. arenata. We also found that V. arenata actively capture and consume more large prey and that M. gracilis more passively capture and consume small prey more reliably. These data suggest that these species have evolved different foraging strategies with V. arenata being risk-prone and M. gracilis being risk-averse. This study provides a more nuanced analysis of niche partitioning between species occupying otherwise similar temporal, habitat, and foraging niches.

Post-traumatic stress disorder (PTSD) is a heterogenous psychiatric condition that affects thousands of individuals each year. Of those who experience this condition, military members including members of the Canadian Armed Forces (CAF) are particularly vulnerable, demonstrating high prevalence rates of PTSD-related symptoms. Moreover, individuals with PTSD are at increased risk for comorbid conditions and are at greater risk for suicide due to the overwhelming, debilitating nature of PTSD symptoms. In previous research, hyperarousal symptoms associated with PTSD have been linked to dysregulation in the locus coeruleus norepinephrine (LC-NE) system, a vast neuromodulatory system responsible for regulating arousal, attention, autonomic and memory-related functions. Advancements in neuroimaging methods have advanced our ability to study connectivity in vivo such that small structures like the LC can be further studied in human samples. Specifically, neuromelanin-sensitive MRI (NM-MRI), a novel, non-invasive neuroimaging method has been shown to detect changes in neuromelanin (NM)-related signal in both the LC and substantia nigra (SN). NM is a dark pigment that accumulates over the lifespan in catecholamine-dominant centers such as the LC and SN and is the by-product of catecholamine oxidation. NM-MRI can be used to image these centers in vivo due to the paramagnetic properties offered by NM. Furthermore, when excess cytosolic catecholamine levels are present in select neurons, NM production is thought to be increased, resulting in increased NM signal from the LC. This could potentially be a marker for dysregulation as many conditions have been associated with variability of this system. Previously, NM-MRI has been used in other clinical settings such as in Parkinson’s disease (PD), Alzheimer’s disease (AD), schizophrenia and depression; however, this current investigation is the first to utilize this imaging modality in the context of PTSD. Specifically, we hypothesized that increased NM-MRI signal in the LC would correlate with increasing severity of hyperarousal symptoms in individuals with PTSD. We also predicted that the opposite would be true for comorbid depression symptom severity, as reduced LC signal has been previously correlated with clinical measures of comorbid depression using NM-MRI. As per our primary hypothesis, we observed a significant positive correlation between NM-MRI signals in the caudal elements of the LC with hyperarousal symptom severity in 22 PTSD subjects (r= 0.54, p= 0.017; partial correlation controlling for depression symptom severity, age, and sex). In contrast, we did not find any evidence to support our secondary hypothesis, because a non-significant trend correlating LC NM-MRI signal and depression symptom severity was obtained (r= -0.30, p=0.22; partial correlation controlling for hyperarousal severity, age, and sex). Based on these results, we were able to build on previously conducted work to further investigate the utility of NM-MRI in the detection of variability in LC-NE system as it pertains to psychiatric conditions known to show dysregulation of this system such as PTSD. In addition, this thesis provides further evidence to support the automation of NM-MRI analytical methods, thus supporting their potential utility for future clinical research. Our findings also provide support for the use of NM-MRI as a potential measure of NE activity; further, this work provided preliminary evidence supporting the use of NM-MRI in a clinical, psychiatric setting, where the technique may serve as a biomarker of PTSD pathology. With these findings in mind, additional validation studies can be conducted to verify the use of NM-MRI as a biomarker for NE system dysregulation. This would potentially allow for advancements in targeted treatment options for PTSD, particularly those targeting the LC-NE system, thus potentially increasing patient stratification and treatment efficacy.

Serious mental illness (SMI) affects 5% of the United States population and is associated with increased morbidity and mortality. Use of high-cost healthcare services is common, including hospitalizations and emergency department (ED) visits. Integrating behavioral and physical healthcare may improve care for consumers with SMI, but prior research findings have been mixed. This quantitative retrospective cohort study addressed the impact of integrated care on physical health and ambulatory care sensitive (ACS) utilization via a program evaluation of an integrated health clinic (IHC) at a community mental health center (CMHC). The research questions assessed whether there was a predictive relationship between IHC enrollment and physical health and ACS-specific service utilization for consumers with SMI when controlling for demographic characteristics and disease severity. Secondary administrative healthcare data, including authorization and electronic medical record data, were provided by the CMHC. Logistic regressions assessed the odds of experiencing an inpatient admission or ED visit before or after IHC enrollment; the predictive relationship between IHC enrollment and service utilization was assessed using multiple linear and Poisson regression analyses. There was no statistically significant impact of integrated care clinic enrollment on physical health or ACS-specific utilization. The sample had lower levels of physical health utilization than would have been expected. In terms of positive social change, results may help the CMHC assess the IHC program, overall clinic success, and use of data. Since policy and payment structures continue to support integrated care models, further research on different programs are encouraged, as each setting and practice pattern is unique.

Une forte exigence de robustesse s’est imposée dans tous les domaines d’application des composants de puissance. Dans ce cadre très contraint, seule une analyse fine des phénomènes liés directement ou indirectement aux défaillances peut garantir une maîtrise de la fiabilité des fonctions assurées par les nouveaux composants de puissance. Cependant, ces phénomènes impliquent des couplages entre des effets électriques, thermiques et mécaniques, rendant leur étude très complexe. Le recours à la modélisation multi-physique bien adaptée s’avère alors déterminant. Dans ce mémoire de thèse, nous proposons une méthodologie de modélisation électrique prenant en compte les effets de la température sur les phénomènes localisés qui initient une défaillance souvent fatale. En prévision de la simulation électrothermique couplée impliquant des transistors MOS de puissance, un modèle électrique thermosensible de ce composant et de sa diode structurelle a été développé. Corrélativement un ensemble de bancs expérimentaux a été mis en œuvre pour l’extraction des paramètres et pour la validation du modèle. Une attention particulière a été accordée à l’étude des phénomènes parasites qui pourraient survenir de manière très localisée suite à une répartition inhomogène de la température et à l’apparition de points chauds. Ainsi les fonctionnements limites en avalanche, avec le déclenchement du transistor bipolaire parasite et de son retournement ont été modélisés. Des bancs spécifiques pour la validation du modèle pour les régimes extrêmes ont été utilisés en prenant des précautions liées à la haute température. Enfin, Le modèle électrique thermosensible complet développé a été utilisé par la société Epsilon ingénierie pour faire des simulations électrothermiques du MOS de puissance en mode d’avalanche en adaptant le logiciel Epsilon-R3D / Strong demand for robustness has emerged in all areas of application of power components.Only a detailed analysis of phenomena related directly or indirectly to failures can ensure thereliability of the functions of the new power components. However, these phenomena involvethe coupling between electrical effects, thermal and mechanical, making their study verycomplex. The use of multi-physics modeling is well suited when determining. In this thesis,we propose a methodology for electrical modeling taking into account the effects of temperatureon the localized phenomena that initiate failure is often fatal. In preparation for thecoupled electro-thermal simulation involving MOS power transistors, an electric thermosensitivemodel of the MOS and its body diode has been developed. Correspondingly a set ofexperimental studies was implemented to extract the parameters and model validation. Particularattention was paid to the study of interference phenomena that could occur in a localizedresponse to an inhomogeneous distribution of temperature and hot spots. Thus the workingslimits avalanche, with the outbreak of parasitic bipolar transistor (snapback) and its reversalwere modeled. Benches specific validations of the model for harsh switching conditions wereused by taking precautions related to high temperature. Finally, the complete thermal electricmodel developed was used by the company “EPSILON Ingénierie” for electro-thermal simulationof power MOS mode Avalanche Software adapting Epsilon-R3D.

No description available.

Online supervised learning from fast-evolving imbalanced data streams has applications in many areas. That is, the development of techniques that are able to handle highly skewed class distributions (or 'class imbalance') is an important area of research in domains such as manufacturing, the environment, and health. Solutions should be able to analyze large repositories in near real-time and provide accurate models to describe rare classes that may appear infrequently or in bursts while continuously accommodating new instances. Although numerous online learning methods have been proposed to handle binary class imbalance, solutions suitable for multi-class streams with varying degrees of imbalance in evolving streams have received limited attention. To address this knowledge gap, the first contribution of this thesis introduces the Online Learning from Imbalanced Multi-Class Streams through Dynamic Sampling (DynaQ) algorithm for learning in such multi-class imbalanced settings. Our approach utilizes a queue-based learning method that dynamically creates an instance queue for each class. The number of instances is balanced by maintaining a queue threshold and removing older samples during training. In addition, new and rare classes are dynamically added to the training process as they appear. Our experimental results confirm a noticeable improvement in minority-class detection and classification performance. A comparative evaluation shows that the DynaQ algorithm outperforms the state-of-the-art approaches. Our second contribution in this thesis focuses on fairness-aware learning from imbalanced streams. Our work is motivated by the observation that the decisions made by online learning algorithms may negatively impact individuals or communities. Indeed, the development of approaches to handle these concerns is an active area of research in the machine learning community. However, most existing methods process the data in offline settings and are not directly suitable for online learning from evolving data streams. Further, these techniques fail to take the effects of class imbalance, on fairness-aware supervised learning into account. In addition, recent fairness-aware online learning supervised learning approaches focus on one sensitive attribute only, which may lead to subgroup discrimination. In a fair classification, the equality of fairness metrics across multiple overlapping groups must be considered simultaneously. In our second contribution, we thus address the combined problem of fairness-aware online learning from imbalanced evolving streams, while considering multiple sensitive attributes. To this end, we introduce the Multi-Sensitive Queue-based Online Fair Learning (MQ-OFL) algorithm, an online fairness-aware approach, which maintains valid and fair models over evolving streams. MQ-OFL changes the training distribution in an online fashion based on both stream imbalance and discriminatory behavior of the model evaluated over the historical stream. We compare our MQ-OFL method with state-of-art studies on real-world datasets and present comparative insights on the performance. Our final contribution focuses on explainability and interpretability in fairness-aware online learning. This research is guided by the concerns raised due to the black-box nature of models, concealing internal logic from users. This lack of transparency poses practical and ethical challenges, particularly when these algorithms make decisions in finance, healthcare, and marketing domains. These systems may introduce biases and prejudices during the learning phase by utilizing complex machine learning algorithms and sensitive data. Consequently, decision models trained on such data may make unfair decisions and it is important to realize such issues before deploying the models. To address this issue, we introduce techniques for interpreting the outcomes of fairness-aware online learning. Through a case study predicting income based on features such as ethnicity, biological sex, age, and education level, we demonstrate how our fairness-aware learning process (MQ-OFL) maintains a balance between accuracy and discrimination trade-off using global and local surrogate models.

Industries such as automotive and industrial automation are pushing the limits of networking technology. The Time-Sensitive Networking (TSN) standards offer a promising solution that can meet the demands of high-bandwidth applications with strict timing constraints, allowing time-critical traffic to coexist with other traffic. However, TSN is relatively new, with much research necessary before it is usable in the industry. This thesis addresses the problem of the common manual configuration of end stations, being time-consuming and error-prone. Through exploring the TSN standards, this thesis attempts to solve the configuration problem by providing a proof-of-concept for both design and implementation of a software architecture managing the end stations and automating their configuration process. Adopting an agile and iterative approach made the complexity of TSN manageable. Furthermore, this thesis has been guided by the research question: How can the configuration of a TSN end station be automated? The design was split into three components: an interface, a configuration manager, and a TSN library. In addition, a communication protocol between the end stations and the Centralized Network Configuration (CNC) is established. The implementation of the proposed design used Python for all three modules in the end station management software, with the REST protocol for the interface. After presenting the implementation, it was evaluated to show the performance of the implemented end station management software. The results showed that the management software would likely not be the bottleneck, as other components it depends on are considerably slower. This thesis and its research contribution offer a practical foundation for continued research and development, such as investigating the configuration of the end stations, providing proofs-of-concept for engineering tools with specific use cases, and finalizing a TSN system. / Industrier som fordons- och industriautomation tänjer på gränserna av nätverksteknik. Time-Sensitive Networking (TSN) standarderna är en lovande lösning som lovar att möta kraven för applikationer med krav på hög bandbredd och strikta tidskrav, som samtidigt tillåter tidskritisk trafik att existera tillsammans med annan trafik. Dock är TSN relativt nytt och i stort behov av mer forskning innan det kan användas i industrin. Den här avhandlingen adresserar problemet med manuell konfiguration av ändstationer, som är både tidskrävande och felbenäget. Genom att utforksa TSN standarderna försöker den här avhandlingen lösa konfigurationsproblemet genom att ta fram ett koncept för design och implementation av en mjukvaruarkitektur för att hantera ändstationer och automatisera deras konfigurationsprocess. Genom att anta en agil och iterativ metod blev komplexiteten hos TSN hanterbar. Dessutom har den här avhandlingen styrts av forskningsfrågan: Hur kan konfigurationen av TSN ändstationer automatiseras? Designen delades upp i tre komponenter: ett gränssnitt, en konfigurationshanterare, samt ett TSN-bibliotek. Utöver detta etablerades ett kommunikationsprotokoll mellan ändstationer och en Centralized Network Configuration (CNC). Implementationen av den föreslagna designen använde sig av Python för de tre komponenterna, med REST-protokollet för gränssnittet. Efter presentationen av implementationen utvärderades den för att visa prestandan hos den implementerade mjukvaran för ändstationen. Resultaten visade att mjukvaran sannolikt inte skulle vara en flaskhals, då andra komponenter som den är beroende av, är betydligt långsammare. Den här avhandligen och dess forskningsbidrag erbjuder en praktiskt grund för fortsatt forskning och utveckling, som undersökning av konfigurationen för ändstationer, framtagning av koncept för ingenjörsverktyg med specifika användningsfall, samt att slutföra ett TSN-system.

Preventive maintenance plays a vital role in optimizing industrial operations. However, detecting equipment needing such maintenance using available data can be particularly challenging due to the class imbalance prevalent in real-world applications. The datasets gathered from equipment sensors primarily consist of records from well-functioning machines, making it difficult to identify those on the brink of failure, which is the main focus of preventive maintenance efforts. In this study, we employ neural network algorithms to address class imbalance and cost sensitivity issues in industrial scenarios for preventive maintenance. Our investigation centers on the "APS Failure in the Scania Trucks Data Set," a binary classification problem exhibiting significant class imbalance and cost sensitivity issues—a common occurrence across various fields. Inspired by image detection techniques, we introduce a novel loss function called Focal loss to traditional neural networks, combined with techniques like Cost-Sensitive Learning and Threshold Calculation to enhance classification accuracy. Our study's novelty is adapting image detection techniques to tackle the class imbalance problem within a binary classification task. Our proposed method demonstrates improvements in addressing the given optimization problem when confronted with these issues, matching or surpassing existing machine learning and deep learning techniques while maintaining computational efficiency. Our results indicate that class imbalance can be addressed without relying on conventional sampling techniques, which typically come at the cost of increased computational cost (oversampling) or loss of critical information (undersampling). In conclusion, our proposed method presents a promising approach for addressing class imbalance and cost sensitivity issues in industrial datasets heavily affected by these phenomena. It contributes to developing preventive maintenance solutions capable of enhancing the efficiency and productivity of industrial operations by detecting machines in need of attention: this discovery process we term predictive maintenance. The artifact produced in this study showcases the utilization of Focal Loss, Cost-Sensitive Learning, and Threshold Calculation to create reliable and effective predictive maintenance solutions for real-world applications. This thesis establishes a method that contributes to the body of knowledge in binary classification within machine learning, specifically addressing the challenges mentioned above. Our research findings have broader implications beyond industrial classification tasks, extending to other fields, such as medical or cybersecurity classification problems. The artifact (code) is at: https://shorturl.at/lsNSY

“Global floods and extreme rainfall events have surged by more than 50% in the past decade and recent studies show that they are occurring four times higher than in 1980” (Neslen, 2018). At the same time, the urban population is rising. Today, 55% of the world’s population lives in urban areas and it is estimated to increase to 70% by 2050 (United Nations, 2018). This expansion of urbanized areas is correlated with the increase of impermeable surfaces that, in case of extreme weather events, are not able to drain the water efficiently. The rainfall-runoff is channelled from roads, parking lots, buildings, and other impervious surfaces to storm drains and sewers that cannot handle the volume. The high ratio of impermeable surfaces and the increased extreme rainfall events cause severe environmental, social, economical problems in urban areas. Merely technical and engineering solutions are no sufficient, therefore a new approach that can maintain and adapt the natural water cycle inside the urban areas is needed. Ecosystem services and resilience thinking have become key principles in adaptation strategies at different levels, from international policies (e.g. Sustainable Development Goals) to local actions (e.g. Copenhagen adaptation plan 2015) and design (e.g. climate resilient San Kjeld in Copenhagen). In this scenario, the design approach of Water Sensitive Urban Design (WSUD) aims to promote resilience at the local level by managing stormwater, encouraging the defence of the aesthetic value of green and blue areas. WSUD is a multidisciplinary approach that involves water management, urban planning, architecture, and landscape design. The main idea of WSUD is that sustainable stormwater systems should be beautiful, meaningful, and educational (Echols, 2007). This master thesis explores the concept of Water Sensitive Urban Design and its application in the cities of Copenhagen, Malmö and Rotterdam. The case study of PHVision in Heidelberg, Germany, is analysed from the concept of WSUD. Design improvements are suggested stemming from the analysed European examples and the theoretical background.