Parmi les facteurs de risque connus de la consommation de substances psychoactives, les traits de personnalité ont été identifiés dans plusieurs études empiriques (Chassin et al., 2009). Toutefois, très peu d’études ont testé le modèle transactionnel (Sameroff, 2009) qui postule qu’il existe des relations bidirectionnelles entre la personnalité et la consommation (Littlefield et al., 2009; Malmberg et al., 2013). Cette étude visait donc à déterminer s’il existe des relations bidirectionnelles significatives entre la consommation et les traits de personnalité durant l’adolescence. L’effet modérateur du genre a été examiné compte tenu des différences de genre connues dans les traits de personnalité et la consommation. Les données utilisées proviennent d’une étude longitudinale prospective de cinq ans chez 1036 adolescents issus de sept écoles de la grande région de Montréal et Québec évalués à trois reprises (première, troisième et cinquième secondaire). Des analyses autorégressives à décalage croisé ont démontré que les traits de personnalité ainsi que la consommation sont stables durant l’adolescence. Chez les filles, les traits de personnalité de l’ouverture, l’extraversion, l’amabilité et le contrôle prédisent la consommation subséquente, tandis que l’amabilité et le contrôle prédisent la consommation ultérieure chez les garçons. En outre, la consommation est significativement associée à la diminution du contrôle et de la stabilité émotionnelle ultérieure chez les garçons et à l’augmentation de la stabilité émotionnelle chez les filles. Une relation bidirectionnelle est observée au niveau du contrôle seulement chez les filles. Des différences selon le genre sont observées pour certaines relations prédictives, mais pas pour la stabilité. / Among the known risk factors of substance use, personality traits have been identified in several empirical studies (Chassin et al., 2009). However, very few studies have tested Sameroff's (2009) transactional model which postulates that there are bidirectional relationships between personality and substance use (Littlefield et al., 2009; Malmberg et al., 2013). This study therefore aimed to determine whether there are significant bidirectional relationships between substance use and personality traits during adolescence. The moderator effect of gender was tested given the known gender differences in personality traits and substance use. The data used came from a five-year prospective longitudinal study of 1036 adolescents from seven schools in the greater Montreal and Quebec areas who were assessed on three occasions (secondary 1, 3 and 5). Autoregressive cross-lagged analyses showed that personality traits and substance use are stable during adolescence. For girls, the personality traits of openness, extraversion, agreeableness, and control predicted subsequent use, while agreeableness and control predicted subsequent use for boys. Furthermore, substance use was significantly associated with subsequent decrease in control and emotional stability in boys and increase in emotional stability in girls. A bidirectional relationship was observed for control only for girls. Gender differences were observed for some predictive relationships, but not for stability.

Bidirectional onboard chargers (OBCs) are becoming mainstream commercial charging equipment for electric vehicles (EVs) because of their compactness, flexibility, and demand-response capabilities for power backup. This thesis focuses on the novel phase shift (PS) modulation techniques for efficiency improvement for bidirectional OBCs, including two-stage onboard chargers (TSOBCs) and single-stage onboard chargers (SSOBCs). A comprehensive overview and investigation of the state-of-the-art solutions of bidirectional OBCs are presented. It reviews the current industrial status, industrial applications, and future trends and challenges. A detailed overview of the promising topologies for bidirectional OBCs, including two-stage and single-stage structures, is also discussed in this thesis. Traditional PS modulation has been widely used in the back-end DC/DC converters of the TSOBCs because of its simple implementation. However, it is challenging to keep high efficiency at boundary operating points within wide specifications. Therefore, to improve efficiency at the boundary point for TSOBCs, the hybrid multiple phase shift (HMPS) modulation technique with minimal peak current optimization is presented to maximize the zero-voltage switching (ZVS) range. Compared to traditional single phase shift (SPS) modulation, the experimental results verify that the presented HMPS modulation strategy provides 1%-2% higher efficiency at the boundary points. On the other hand, an improved compact SSOBC topology and novel PS modulation techniques are proposed. Since the traditional PS modulation is challenging for AC/DC converters to keep a unity power factor (PF), novel PS modulation techniques are presented for the proposed SSOBC. Firstly, a sinusoidal single phase shift (SSPS) modulation introduces a sinusoidal phase shift to maintain a high PF and high efficiency within a wide operating point. However, due to the high current at the zero-crossing point of the grid voltage of the SSPS modulation, the novel adaptive sinusoidal single phase shift (ASSPS) modulation is presented to address this issue, which reduces conduction loss and increases efficiency. Secondly, based on the ASSPS modulation, the adaptive sinusoidal extended phase shift (ASEPS) modulation with minimal peak current optimization is presented to introduce one more degree of freedom to extend the ZVS flexibility, which reduces switching loss. Moreover, the minimal peak current optimization reduces transformer current, further decreasing conduction losses. Therefore, the power loss is minimized. Finally, this thesis presents the general design guideline of a 6 kW Silicon Carbide (SiC)-based bidirectional SSOBC, contributing to the further development of bidirectional SSOBC application. Experimental results verify the operating principle and high PF of the proposed SSPS, ASSPS, and ASEPS modulation. 1 kW experimental testing has validated that the peak efficiency is 95.3% with ASSPS modulation and 95.9% with ASEPS modulation. Compared to the existing pulse width modulation (PWM), the ASSPS modulation increased efficiency by 1.1%, and ASEPS modulation further increased by 1.7%. / Thesis / Doctor of Philosophy (PhD)

For music genre classification purposes, the importance of an intelligent and content-based selection of audio samples has been mostly overlooked. One common approach toward representative results is to select samples at predetermined locations. This is done to avoid analysis of the full audio during classification. While methods in music thumbnailing could be used to find representative samples for genre classification, it has not yet been demonstrated. This thesis showed that efficient and genre representative sampling can be performed with a machine learning model (bidirectional RNN with either LSTM or GRU cells). The model was trained using a sub-optimal genre classifier and computationally inexpensive audio features. The genre classifier was used to compute losses for evenly spaced samples in 14000 tracks. The losses were then used as targets during training. Root mean square energy and zero-crossing rate were used as features, computed over relatively large time steps and wide intervals. The proposed framework can be used to give better predictions with trained genre classifiers and most likely also train, or retrain, them for higher classification accuracy at a low computational cost. / Vid musikgenreklassificering har betydelsen av ett intelligent och innehållsbaserat urval allt som oftast förbisetts. En ansats till ett representativt resultat görs vanligtvis genom att ett antal kortare utdrag tas vid förutbestämda tidpunkter. Detta görs för att under en klassificering undvika att analysera hela musikverket. Fastän det existerar metoder inom music thumbnailing för att hitta representativa urval har de ännu inte tillämpats inom genreklassificering. I denna uppsats visades att ett effektivt och genrerepresentativt musikurval kan utföras med en maskininlärningsmodell (dubbelriktad RNN med antingen LSTM- eller GRU-celler). Modellen tränades med hjälp av en suboptimal genreklassificerare och beräkningsmässigt enkla ljudattribut. Genreklassificeraren användes för att beräkna förlusten av jämnt fördelade urval i 14000 musikverk. Förlusterna användes sedan som utdata under träningen. Kvadratiskt energimedelvärde och zero-crossing rate beräknades över relativt långa tidssteg och breda intervall och användes som indata. Det föreslagna ramverket kan till beräkningsmässigt låga kostnader användas för att ge bättre förutsägelser med redan tränade genreklassificerare och sannolikt träna, eller omträna, dessa för högre noggrannhet vid klassificering.

This research presents an innovative approach to extracting information from Positive Energy Districts (PEDs), urban areas generating surplus energy. PEDs are integral to the European Commission's SET Plan, tackling housing challenges arising from population growth. The study refines BERT to categorize PED-related entities, producing a cutting-edge NER model and an integrated pipeline of diverse NER tools and data sources. The model achieves an accuracy of 0.81 and an F1 Score of 0.55 with notably high confidence scores through pipeline evaluations, confirming its practical applicability. While the F1 score falls short of expectations, this pioneering exploration in PED information extraction sets the stage for future refinements and studies, promising enhanced methodologies and impactful outcomes in this dynamic field. This research advances NER processes for Positive Energy Districts, supporting their development and implementation.

No description available.

No description available.

No / This paper presents a new partitioning strategy with the objective of increasing scalability by reducing computational effort of routing in networks. The original network is partitioned into blocks (subnetworks) so that there is a bi-directional link between any two blocks. When there is a connection request between a pair of nodes, if the nodes are in the same block, we only use the small single block to derive routings. Otherwise we combine the two blocks where the two nodes locate and in this way the whole network will never be used. The strategy is generic in that it can be used in any underlying routing algorithms in the network layer and can be applied to any networks with fixed topology such as fixed wired subnetworks of the Internet. The performance of this strategy has been investigated by building a simulator in Java and a comparison with existing stochastic partitioning techniques is shown to give superior performance in terms of trade-off in blocking probability (the probability of failure to find a path between source and destination satisfying QoS constraints) and reduction of computational effort.

Detta examensarbete fokuserar på utveckling och utvärdering av en ny metod för att räkna dubbelriktade personflöden inomhus med hjälp av Ultrasonic Time-of-Flight teknik. Projektets huvudsyfte är att skapa en kostnadseffektiv, strömsnål och integritetsvänlig lösning som är i linje med lagar som GDPR. Studien börjar med en kort genomgång av tillgängliga tekniker för personflödesmätning, men det blir tydligt att dessa tekniker brister i kraven för den önskade tekniken. Mot denna bakgrund framstår Ultrasonic Time-of-Flight som en lovande kandidat på grund av sin förmåga att detektera objekt och rörelseriktningar utan att samla in personligt identifierbar information. För att realisera detta projekt har ett utvecklingskit baserat på sensorn CH201 från Chirp Microsystems använts. Sensorns låga strömförbrukning och förmåga att mäta avstånd i ett brett synfält är det som är lovande i tekniken. Ett akustiskt hölje optimerar sensorernas synfält och minimerar störningar. Experimentdelen av arbetet inkluderar uppbyggnaden av en testmiljö där sensorernas förmåga att korrekt räkna individer och bestämma deras rörelseriktning testas. Resultaten från dessa tester visar på hög noggrannhet i detektering av enskilda individer som passerar, men har lägre noggrannhet då flera personer passerar samtidigt. Vidare diskuteras potentialen för att vidareutveckla systemet för att även kunna hantera större personflöden och mer komplexa scenarion, som flera personer som rör sig bredvid varandra i olika riktningar. En kritisk granskning av systemets prestanda under längre tidsperioder och i olika miljöer föreslås som framtida forskningsarbete för att ytterligare validera och förbättra tekniken. Sammanfattningsvis demonstrerar detta arbete potentialen hos tekniken som en säker och integritetsvänlig lösning för effektiv övervakning av personflöden. Med ytterligare utveckling och anpassning förväntas tekniken kunna uppfylla en ännu högre noggrannhet. / This thesis focuses on the development and evaluation of a new method for measuring bidirectional indoor people flows using Ultrasonic Time-of-Flight technology. The main purpose of the project is to create a cost-effective, low-power, and privacy-friendly solution that complies with laws like the GDPR. It begins with a short review of existing techniques for measuring people flow, concluding that these technologies do not support the goal of the new technology. Ultrasonic Time-of-Flight emerges as a promising candidate due to its ability to detect objects and directions of movement without collecting personally identifiable information. To realize this project, a development kit based on the CH201 sensor from ChirpMicrosystems has been used. The sensor’s low power consumption and ability to measure distances in a wide field of view are what made the technology promising. An acoustic enclosure optimizes the sensors’ field of view and minimizes interference. The experimental part of the work includes the construction of a test environment where the sensors’ ability to accurately count individuals and determine their direction of movement is tested. The results from these tests show high accuracy in detecting individual passersby but encounter more problems with multiple individuals simultaneously. Further discussions will explore the potential for developing the system to manage larger crowds and more complex scenarios, such as multiple people moving side by side in different directions. A critical review of the system’s performance over longer periods and in different environments is proposed as future research work to further validate and improve the technology. In conclusion, this work demonstrates the potential of the technology as a secure and privacy-friendly solution for effective monitoring of people flows. With further development and adaptation, the technology is expected to offer significantly better accuracy.

<p>The numerical simulation of physical systems plays a key role in different fields of science and engineering. The popularity of numerical methods stems from their ability to simulate complex physical phenomena for which analytical solutions are only possible for limited combinations of geometry, boundary, and initial conditions. Despite their flexibility, the computational demand of classical numerical methods quickly escalates as the size and complexity of the model increase. To address this limitation, and motivated by the unprecedented success of Deep Learning (DL) in computer vision, researchers started exploring the possibility of developing computationally efficient DL-based algorithms to simulate the response of complex systems. To date, DL techniques have been shown to be effective in simulating certain physical systems. However, their practical application faces an important common constraint: trained DL models are limited to a predefined set of configurations. Any change to the system configuration (e.g., changes to the domain size or boundary conditions) entails updating the underlying architecture and retraining the model. It follows that existing DL-based simulation approaches lack the flexibility offered by classical numerical methods. An important constraint that severely hinders the widespread application of these approaches to the simulation of physical systems.</p> <p><br></p> <p>In an effort to address this limitation, this dissertation explores DL models capable of combining the conceptual flexibility typical of a numerical approach for structural analysis, the finite element method, with the remarkable computational efficiency of trained neural networks. Specifically, this dissertation introduces the novel concept of <em>“Finite Element Network Analysis”</em> (FENA), a physics-informed, DL-based computational framework for the simulation of physical systems. FENA leverages the unique transfer knowledge property of bidirectional recurrent neural networks to provide a uniquely powerful and flexible computing platform. In FENA, each class of physical systems (for example, structural elements such as beams and plates) is represented by a set of surrogate DL-based models. All classes of surrogate models are pre-trained and available in a library, analogous to the finite element method, alleviating the need for repeated retraining. Another remarkable characteristic of FENA is the ability to simulate assemblies built by combining pre-trained networks that serve as surrogate models of different components of physical systems, a functionality that is key to modeling multicomponent physical systems. The ability to assemble pre-trained network models, dubbed <em>network concatenation</em>, places FENA in a new category of DL-based computational platforms because, unlike existing DL-based techniques, FENA does not require <em>ad hoc</em> training for problem-specific conditions.</p> <p><br></p> <p>While FENA is highly general in nature, this work focuses primarily on the development of linear and nonlinear static simulation capabilities of a variety of fundamental structural elements as a benchmark to demonstrate FENA's capabilities. Specifically, FENA is applied to linear elastic rods, slender beams, and thin plates. Then, the concept of concatenation is utilized to simulate multicomponent structures composed of beams and plate assemblies (stiffened panels). The capacity of FENA to model nonlinear systems is also shown by further applying it to nonlinear problems consisting in the simulation of geometrically nonlinear elastic beams and plastic deformation of aluminum beams, an extension that became possible thanks to the flexibility of FENA and the intrinsic nonlinearity of neural networks. The application of FENA to time-transient simulations is also presented, providing the foundation for linear time-transient simulations of homogeneous and inhomogeneous systems. Specifically, the concepts of Super Finite Network Element (SFNE) and network concatenation in time are introduced. The proposed concepts enable training SFNEs based on data available in a limited time frame and then using the trained SFNEs to simulate the system evolution beyond the initial time window characteristic of the training dataset. To showcase the effectiveness and versatility of the introduced concepts, they are applied to the transient simulation of homogeneous rods and inhomogeneous beams. In each case, the framework is validated by direct comparison against the solutions available from analytical methods or traditional finite element analysis. Results indicate that FENA can provide highly accurate solutions, with relative errors below 2 % for the cases presented in this work and a clear computational advantage over traditional numerical solution methods. </p> <p><br></p> <p>The consistency of the performance across diverse problem settings substantiates the adaptability and versatility of FENA. It is expected that, although the framework is illustrated and numerically validated only for selected classes of structures, the framework could potentially be extended to a broad spectrum of structural and multiphysics applications relevant to computational science.</p>

In order to increase the power density, the discontinuous conducting mode (DCM) and small inductance is adopted for high power bidirectional dc-dc converter. The DCM related current ripple is minimized with multiphase interleaved operation. The turn-off loss caused by the DCM induced high peak current is reduced by snubber capacitor. The energy stored in the capacitor needs to be discharged before device is turned on. A complementary gating signal control scheme is employed to turn on the non-active switch helping discharge the capacitor and diverting the current into the anti-paralleled diode of the active switch. This realizes the zero voltage resonant transition (ZVRT) of main switches. This scheme also eliminates the parasitic ringing in inductor current. This work proposes an inductance and snubber capacitor optimization methodology. The inductor volume index and the inductor valley current are suggested as the optimization method for small volume and the realization of ZVRT. The proposed capacitance optimization method is based on a series of experiments for minimum overall switching loss. According to the suggested design optimization, a high power density hardware prototype is constructed and tested. The experimental results are provided, and the proposed design approach is verified. In this dissertation, a general-purposed power stage model is proposed based on complementary gating signal control scheme and derived with space-state averaging method. The model features a third-order system, from which a second-order model with resistive load on one side can be derived and a first-order model with a voltage source on both sides can be derived. This model sets up a basis for the unified controller design and optimization. The Δ-type model of coupled inductor is introduced and simplified to provide a more clearly physical meaning for design and dynamic analysis. These models have been validated by the Simplis ac analysis simulation. For power flow control, a unified controller concept is proposed based on the derived general-purposed power stage model. The proposed unified controller enables smooth bidirectional current flow. Controller is implemented with digital signal processing (DSP) for experimental verification. The inductor current is selected as feedback signal in resistive load, and the output current is selected as feedback signal in battery load. Load step and power flow step control tests are conducted for resistive load and battery load separately. The results indicate that the selected sensing signal can produce an accurate and fast enough feedback signal. Experimental results show that the transition between charging and discharging is very smooth, and there is no overshoot or undershoot transient. It presents a seamless transition for bidirectional current flow. The smooth transition should be attributed to the use of the complementary gating signal control scheme and the proposed unified controller. System simulations are made, and the results are provided. The test results have a good agreement with system simulation results, and the unified controller performs as expected. / Ph. D.