Search

This research aims to assist farmers in controlling plant pests, birds and insect, by applying an appropriate electronic technology. The system operates in an automatic and remote-controlled modes using IoT technology. We used a solar powered systems that are environmentally friendly with zero carbon emissions, as an energy source for the systems. The test results showed that the bird pest control system worked well according to the automatic and remote-control modes design. We compare those two methods to evaluate the most effective between those two methods. Bird pest control using two actuators, sound and movement, outsmart control using only sound by success ratio 70:50. The insect pest control system worked well with the remote-control mode through Telegram application on farmers’ smartphones.

BACKGROUND: Reference to an intrinsic healing mechanism or an inner healer is commonplace amongst psychedelic drug-using cultures. The inner healer refers to the belief that psychedelic compounds, plants or concoctions have an intrinsically regenerative action on the mind and brain, analogous to intrinsic healing mechanisms within the physical body, for example, after sickness or injury. AIMS: Here, we sought to test and critique this idea by devising a single subjective rating item pertaining to perceived inner healing effects. METHODS: The item was issued to 59 patients after a single high (25 mg, n = 30) or placebo (1 mg, n = 29) dose of psilocybin in a double-blind randomised controlled trial of psilocybin for depression. RESULTS: Inner healer scores were higher after the high versus placebo dose of psilocybin (t = 3.88, p < 0.001). Within the high-dose sub-sample only, inner healer scores predicted improved depressive symptomatology at 2 weeks post-dosing. CONCLUSIONS: The principle of activating inner healing mechanisms via psychedelics is scientifically nascent; however, this study takes a positivist and pragmatic step forward, asking whether it warrants further examination.

Títol alternatiu emprat a la intranet docent de l'EPSEVG: "Dispensador automàtic d'aerosol amb propulsor natural", L'objectiu principal d'aquest projecte és millorar un dispensador d’aerosol automàtic per fer-lo més respectuós amb el medi ambient i ampliar el seu públic objectiu. En la comparació de diferents models d'aerosols, s'ha observat que tots són recarregables i tenen un preu semblant, però l'aspecte rellevant és el seu funcionament intern i els components de la fragància. Per millorar la sostenibilitat del producte, s’ha optat per utilitzar propel·lents d’origen natural i reformular la fragància. No obstant això, aquest canvi ha afectat el rendiment del dispositiu, provocant una disminució de la pressió interna a mesura que el refill es va buidant. La investigació se centrarà en estudiar com mantenir l'eficiència del dispositiu malgrat aquest canvi en la formulació de la fragància i complir les normatives mediambientals. Per això, es realitzen diversos assajos per determinar la millor combinació de formulació de la fragància i de propel·lent, així com estudiar com aquest canvi afecta el funcionament del dispositiu. També explorarem com aconseguir una polvorització uniforme de la fragància, tenint en compte els canvis en la formulació i en el mecanisme d'aspersió. Finalment s’extreuen les conclusions que confirmen si la producció d’aquest nou dispositiu es viable, en aquest cas caldrà realitzar més assajos ja que el temps i els recursos han sigut limitats. En resum, el nostre objectiu és desenvolupar un dispensador d’aerosol més sostenible i eficient, que compleixi amb les normatives mediambientals i mantingui la seva competitivitat en el mercat., The main objective of this project is to improve an automatic aerosol dispenser to make it more environmentally friendly and to expand its target audience. In comparing different aerosol models, it has been observed that all are rechargeable and have a similar price, but the relevant aspect is their internal operation and the components of the fragrance. To improve the product's sustainability, it was decided to use naturally derived propellants and reformulate the fragrance. However, this change has affected the device's performance, causing a decrease in internal pressure as the refill empties. The research will focus on studying how to maintain the device's efficiency despite this change in fragrance formulation and comply with environmental regulations. Therefore, various tests are being conducted to determine the best combination of fragrance formulation and propellant, as well as to study how this change affects the device's operation. We will also explore how to achieve a uniform spray of the fragrance, considering the changes in formulation and the spraying mechanism. Finally, conclusions will be drawn to confirm whether the production of this new device is viable. If so, further testing will be required as time and resources have been limited. In summary, our goal is to develop a more sustainable and efficient aerosol dispenser that meets environmental regulations and maintains its competitiveness in the market., Outgoing

Risk management has surged in importance across companies of all sizes due to escalating cyberthreats and as sociated costs. This impact disproportionately affects smaller enterprises due to resource constraints and limited technological expertise. This study explores diverse risk management frame works applicable to various business domains and proposes an open-source, resource-efficient solution leveraging containeriza tion for virtualization. Our solution identifies both wired and wireless network devices, offering qualitative and quantitative risk assessment reports over time, along with mitigation options. A proof of concept conducted on a simulated medium-sized enterprise (SME) showcased the detection and assessment of network devices over 20 business days, illustrating risk evolution and escalation in the absence of proactive management.

Accessibility in web development is a subject which has garnered increasing attention over the last few years due to the introduction of heightened compliance laws. Occurring during a time of rapid growth for commercial generative artificial intelligence, many tools that assist in the web development process, such as generative code tools for front end development, have been created. This paper sets out to evaluate some of these tools from an accessibility standpoint: is the code generated by these tools accessible? How can these tools be utilized to facilitate the generation of accessible code? The goal was to get a detailed view of the level of accessibility possible for different levels of effort in using generative code tools. Testing was conducted on code generated by two prototype-to-code plugins on a web prototype in various stages of completeness, using the generative code tools to various levels of advanced functionality. The results show a lack of accessibility in all generated code, regardless of best practices in the prototype or usage of advanced features in the generative code tools themselves. Using automatic checkers also proved to be an insufficient way to check automatically generated code, as several false positives caused by poor code flew under the radar. However, the tools are capable of solving some simpler problems by themselves but require further development to live up to current standards of both web development and accessibility. Further testing on a smaller scale could yield deeper insight into how to best leverage these tools to solve more specific problems.

Producción Científica, The structural characterization is an essential task in the study of porous materials. To achieve reliable results, it requires to evaluate images with hundreds of pores. Current methods require large time amounts and are subjected to human errors and subjectivity. A completely automatic tool would not only speed up the process but also enhance its reliability and reproducibility. Therefore, the main objective of this article is the study of a deep-learning-based technique for the structural characterization of porous materials, through the use of a convolutional neural network. Several fine-tuned Mask R–CNN models are evaluated using different training configurations in four separate datasets each composed of numerous SEM images of diverse polymeric porous materials: closed-pore extruded polystyrene (XPS), polyurethane (PU), and poly(methyl methacrylate) (PMMA), and open-pore PU. Results prove the tool capable of providing very accurate results, equivalent to those achieved by time-consuming manual methods, in a matter of seconds., MCIN/AEI/10.13039/501100011033 and the EU NextGenerationEU/PRTR program (PLEC2021-007705), Junta de Castilla y León y el programa EU-FEDER (CLU-2019-04)

Precision viticulture is increasingly being applied to automate and optimize grape production in the vineyard. This paper describes the development of a method for automatic selection of regions of interest from hyperspectral images obtained of a row of vines and intended for prediction of soluble solids content. For this purpose, a dataset consisting of hyperspectral images of a row of ‘Sangiovese’ wine grapes was adopted. Hyperspectral images were acquired directly in the field by means of a hyperspectral imaging Vis/NIR system (400–1000 nm) mounted on a ground-based vehicle. The analyses were carried out on 17 different days, under clear or partly cloudy conditions, in the period between post-veraison and harvest. The vineyard row of Sangiovese vines was divided into 11 sections and a hyperspectral image for each section for each day of analysis was acquired. The regions of interest of the hyperspectral images, comprising the areas representing the grapes, were selected using a PLS-DA-based method. The best PLS-DA model provided excellent results, with sensitivity and specificity values of 0.991 and 0.996, respectively. The mean spectra of the selected regions of interest (ROI) were finally used to predict the soluble solids content (SSC) of the grapes by PLS regression to a primary reference analysis. The results of SSC predictions using the automatic selection of ROIs (R2CV = 0.74 and RMSECV = 0.86 °Brix) were on par with similar regression based on carefully manual selection of ROIs (R2CV = 0.73 and RMSECV = 0.87 °Brix).

Misalignment of the rotation axis causes severe artifacts in X-ray computed tomography. Calibration of this parameter is often insufficient for sub-micron resolution measurements and needs to be corrected during the post-processing. This correction can be accelerated by various automatic methods. These vary in mechanisms and performance, making them suitable for different use-cases. This work summarizes existing automatic methods for estimating the rotation axis in X-ray computed tomography, with a focus on sub-micron applications. Some of the methods are implemented and compared in the context of a laboratory sub-micron scanner to demonstrate practical considerations of this task.

Our emotions and sentiments are influenced by naturalistic stimuli such as the movies we watch and the songs we listen to, accompanied by changes in our brain activation. Comprehension of these brain-activation dynamics can assist in identification of any associated neurological condition such as stress and depression, leading towards making informed decision about suitable stimuli. A large number of open-access functional magnetic resonance imaging (fMRI) datasets collected under natzuralistic conditions can be used for classification/prediction studies. However, these datasets do not provide emotion/sentiment labels, which limits their use in supervised learning studies. Manual labeling by subjects can generate these labels, however, this method is subjective and biased. In this study, we are proposing another approach of generating automatic labels from the naturalistic stimulus itself. We are using sentiment analyzers (VADER, TextBlob, and Flair) from natural language processing to generate labels using movie subtitles. Subtitles generated labels are used as the class labels for positive, negative, and neutral sentiments for classification of brain fMRI images. Support vector machine, random forest, decision tree, and deep neural network classifiers are used. We are getting reasonably good classification accuracy (42-84%) for imbalanced data, which is increased (55-99%) for balanced data., Naše emoce a pocity jsou ovlivňovány přirozenými podněty, jako jsou filmy, které sledujeme, a písně, které posloucháme, což je doprovázeno změnami v aktivaci našeho mozku. Pochopení dynamiky této mozkové aktivace může pomoci při identifikaci jakéhokoli souvisejícího neurologického stavu, jako je stres a deprese, což vede k informovanému rozhodování o vhodných podnětech. Pro klasifikační/predikční studie lze využít velké množství volně přístupných souborů dat ze zobrazování funkční magnetickou rezonancí (fMRI) shromážděných za naturalistických podmínek. Tyto datové soubory však neposkytují označení emocí/sentimentů, což omezuje jejich použití ve studiích učení pod dohledem. Ruční označování subjekty může tyto štítky generovat, nicméně tato metoda je subjektivní a zkreslená. V této studii navrhujeme jiný přístup generování automatických štítků ze samotných naturalistických podnětů. Ke generování štítků pomocí filmových titulků používáme analyzátory sentimentu (VADER, TextBlob a Flair) ze zpracování přirozeného jazyka. Štítky vygenerované z titulků se používají jako štítky tříd pro pozitivní, negativní a neutrální sentimenty pro klasifikaci snímků fMRI mozku. Použity jsou klasifikátory typu podpůrný vektor, náhodný les, rozhodovací strom a hluboká neuronová síť. Pro nevyvážená data dosahujeme poměrně dobré klasifikační přesnosti (42-84 %), která se pro vyvážená data zvyšuje (55-99 %).

Accumulation of waste is a major global concern, and recycling is considered one of the most effective methods to solve the problem. However, recycling requires proper segregation of waste according to waste types. This paper develops an automatic waste segregator, capable of identifying and segregating six types of wastes; metal, paper, plastic, glass, cardboard, and others. The proposed system employs Convolutional Neural Network (CNN) technology, specifically the Inception-v3 architecture, as well as two physical sensors; weight and metal sensors, to classify and segregate the waste. Overall classification accuracy of the system is 86.7%. Classification performance of the developed waste segregator has been evaluated further using the precision and recall; with high precision obtained for cardboard, metal, and other waste types, and high recall for metal and glass. These results demonstrate the applicability of the developed system in effectively segregating waste at source, and thereby, reducing the need for the commonly labor-intensive segregation at waste facility. Deploying the system has the potential of reducing waste management problems by assisting recycling companies in sorting recyclable waste, through automation.

The last decade has seen a rapid advancement in the fields of artificial intelligence (AI) and automation, with terms such as machine learning, deep learning and algorithms finding their way into everyday conversations. The usage of AI is evergrowing and present in many areas of life such as health care, financial services, and social media interactions. One subject which has garnered a significant amount of attention is the introduction of autonomous vehicles (AVs) and their promise of a safer and more efficient global traffic environment. However, while self-driving technology is developing at an exponential rate, associated laws and regulations are falling behind. The legal issues connected to AVs tend to differ from one jurisdiction to another, but a concern shared among most is the uncertain state of the legal landscape with respect to criminal liability. This study is focused on critically evaluating this uncertain legal landscape from a Swedish perspective. The method used in this thesis is a blend of legal dogmatics and legal analytics. A specific branch of legal analytics, legal informatics, is largely applied due to its suitability in analyzing the relationship between law and information technology. After a brief introduction to AI and AVs, the investigation begins by examining current law to determine how a general implementation of AVs would be handled by existing regulation. The thesis then continues to analyze the potential effects of the European proposal of an electronic legal personhood for AI and its associated ethical considerations. Lastly, the paper explores the Swedish proposal for a specialized AV regulation in comparison to current Swedish and international law. The conclusion of this study is that certain traffic scenarios tend to make the criminal liability assessment clearer than others. A human driver will generally not be held responsible if a vehicle driven by an automated driving system causes an incident. In most cases, the liability

Street view imagery (SVI) is one of the largest (growing) resources in urban analytics. A global close-up of the urban environment, if you will, which is rich in (untapped) information such as floor count. Floor count is useful in many applications, from improving energy consumption calculations to creation of 3D city models without elevation data. So far, efforts to extract floor count from SVI are mainly approached as a classification problem with the use of convolutional neural networks (CNNs). Limitations of this approach include the need of large (manually annotated) datasets, and uncertainty how these models learn to count storeys. Therefore, we aim to develop a method that can be trained on available datasets and determine floor count in a more explainable manner. In order to make the floor count determination method more transparent, we mimic the row-wise counting of storeys as humans do: by vertically parsing a column of windows (and occasional door). Façade parsing is a common computer vision task that we can solve with deep learning. In this work, we employ the Mask R-CNN framework, that is trained on publicly available datasets, for the detection and segmentation of windows and doors. Then, the vertical distribution of detected / segmented windows and doors is estimated by computing the kernel density estimation function. The floor count is extracted by finding the number of maxima in the function, as the maxima represent the dense areas of windows and doors on a horizontal axis (i.e. storeys). To improve the results, an automatic image rectification is added as pre-processing step that enforces the regularity and repetitive occurrence of windows and doors. The full pipeline thus consists of three stages: 1) automatic image rectification, 2) window and door detection/ segmentation with Mask RCNN, 3) floor count estimation via maxima finding on the kernel density estimation (KDE) function. In addition, a small "wild" dataset was created that cont, Geomatics

Introduktion: Blodtryck är en livsviktig funktion som bidrar till syretillförseln till kroppens olika organ. Blodtryck går att mäta och är en viktig indikation på det kardiovaskulära systemets funktion. Det finns olika metoder att mäta blodtryck, bland annat auskultatoriskt, automatiskt och med hjälp av Dopplerteknik. Syftet var att jämföra de tre olika metoderna för att se om det förelåg en signifikant skillnad mellan metoderna. Material och metod: Studien bestod av 30 deltagare. Studiedeltagaren vilade i tio minuter före första blodtrycksmätningen, därefter mättes deras blodtryck i en slumpad ordning med auskultatorisk, automatisk och med Dopplermetod. För att kontrollera om det förelåg en statistisk signifikant skillnad användes en variansanalys (ANOVA) och ett parat t-test. Resultat och slutats: Det förelåg ingen signifikant skillnad mellan de systoliska blodtrycken. Däremot mellan de diastoliska trycken registrerat med den auskultatoriska och automatiska metoden förelåg det en signifikant skillnad, men den anses inte ha en klinisk betydelse., Introduction: Blood pressure is a vital function that contributes to the supply of oxygen to the body’s various organs. Blood pressure can be measured and is an important indicator of the cardiovascular system function. There are different methods to measure blood pressure, these are, for example auscultatory, automatic and with the help of a Doppler technique. The aim was to compare the three different methods and see if there was a significant difference between them. Material and method: The study consisted of 30 participants. The participant rested for ten minutes before the first examination and then their blood pressure was measured in a random order by the three methods: auscultatory, automatic and the Doppler method. To process the collected data an analysis of variance test (ANOVA) and a paired t-test were used. Results and conclusion: There were no significant difference between the systolic blood pressures. In contrast, there was a difference between the diastolic blood pressures registered with the auscultatory and automatic methods, but this was not considered to be of clinical significance.

Traditional methods of watering plants based on timers have been shown to be wasteful when compared to sensor based watering. While sensor based irrigation systems do exist, current implementations are both costly and needlessly complex. Our goal was to design an irrigation sensor system is cheap, easy to implement, and saves more water than traditional watering systems. Our approach was to create a sensor node network that is both reliable and cost effective when installing it. With this approach we were able to create a dependable tree like sensor network that could read moisture data from soil and have it visualized in an Android application. In regards to our Android application, we were able to implement features such as sensor calibration in order to obtain accurate readings. Our tree like network was also able to be consistent in being able to send data packets back and forth with each sensor having a reach of up to 100 meters and being able to have a total of 3125 radio modules in use. In conclusion we were able to create a reliable and cost effective irrigation system that could be used for both farms and home gardens. We were able to meet our initial goals which were to create an user friendly application that could be used to read and visualize data as well creating a sensor network that is low-power and is low cost.  Advisor: Professor Sitao Huang, Department of Electrical Engineering and Computer Science

This thesis proposes an algorithm for automatically detecting and classifying participants into their respective teams in sports events using computer vision and artificial intelligence techniques. The algorithm is designed as a pipeline with four main stages: object detection, feature extraction, tracking and team classification stages. We create a manually annotated dataset and train a feature extractor based on a Siamese network. The feature extractor learns to extract discriminant characteristics from players. Finally, the team classification stage utilizes clustering and classifying techniques on the extracted features. Our algorithm demonstrates accurate team recognition. The algorithm's outputs enhance automated match production, player tracking, highlight generation, and real-time statistics. It also provides valuable insights for coaches, analysts, and sports enthusiasts. This research contributes to the field of sports broadcasting and analytics, enhancing the viewing experience and enabling data-driven decision-making in sports.

Automatic pronunciation assessment programs allow learners to practice pronunciation independently and receive immediate feedback based on acoustic analysis of their speech compared to a pronunciation model. The programs display a high level of reliability when assessing speakers’ pronunciation accuracy. The paper aims to present the results of an investigation into a recent experimental application of an automatic pronunciation assessment online program used within the flipped classroom of English phonetics and phonology course taught to foreign language pre-service teachers of English. Flipped learning approach was selected to offload the preliminary activities and tasks that do not require the immediate attention of the teacher to an online learning management system. Thus, course participants’ face-to-face learning could be enhanced and more effective. The participants could practice and receive tailored feedback from an online program before the face-to-face class. After the individual pronunciation practice, they could discuss more complex questions with the lecturer and, in collaboration with fellow students. The paper will present the survey results among the course participants who evaluated their experience with automatic pronunciation evaluation. Special attention is paid to the potential use of such programs for developing linguistic competence.

Analyzing food consumption patterns can provide valuable insights into the development of obesity and eating disorders. The detection and quantification of chewing strokes are essential to facilitate this analysis. One approach to food intake analysis involves evaluating chewing sounds generated during the eating process. These sounds were recorded by microphones placed to the user’s outer ear canal. Aside from discovering the most accurate solution, the algorithms used must demonstrate sufficient efficiency to operate on low-power embedded ear-worn hardware. Three algorithms for automated chewing detection were evaluated with the help of two datasets. The first dataset consists of the food intake sounds from the consumption of three types of food. The second dataset consists of environmental noise. The data were manually labeled by recognizing mastication sounds’ visual and audio characteristics. Precision of over 80% was achieved by all algorithms in the dataset consisting of only chewing sounds. Finally, an efficient solution has been developed to distinguish between speech and chewing sounds., CSE3000 Research Project, Computer Science and Engineering

Today, manufacturing factory layout setups are most often manually designed and kept up-to-date during their lifecycle with computer-aided design software’s, so that analyses, verification simulations, and decisions can continuously be done. This manual approach is a cumbersome iterative process to collect the necessary information, with a high risk of faulty inputs and updates. Often the virtual descriptions do not match the physical version of the factory setup. This research presents a digital twin solution where physical equipment is connected to a virtual representation of the same equipment and automatically updates the virtual environment with the spatial position of the physical equipment and a proposed way to evaluate it. The physical equipment either has inbuilt sensors or has been equipped with external wireless sensors to track the spatial position. The metadata are distributed via Node-RED (a tool to visualize Internet of Things) to the simulation software Industrial Path Solutions, where the virtual equipment is repositioned based on data from the physical equipment. The result shows that it is possible to send spatial position information from a physical equipment and update the corresponding virtual description of the equipment in its virtual environment. The accuracy of the updates has been evaluated with manual measurements. Hence, the virtual environment of the factory setup, i.e., the digital twin, updates automatically based on the data sent by the physical equipment. With an up-to-date virtual environment, more accurate simulations and optimizations of the factory setup can be achieved. Examples of such possibilities are to evaluate ergonomic conditions or to optimize robot paths for robot cells in realistic and up-to-date virtual environments. Also, one could imagine making virtual reality visits to hazardous or sensitive factory environments in a safe way or studying things that otherwise would be hard or impossible in the real world., CC BY-NC 4.0Andreas Lind [andreas.lind@scania.com]This research was realized with the support of Scania CV AB’s Global Industrial Development and the VF-KDO profile (Virtual Factories with Knowledge-Driven Optimization) at the University of Skövde, funded by the Knowledge Foundation.

This dissertation introduces a biologically-detailed computational model of how rule-guided behaviors become automatic. The model assumes that initially, rule-guided behaviors are controlled by a distributed neural network centered in the prefrontal cortex, and that in addition to initiating behavior, this network also trains a faster and more direct network that includes projections from sensory association cortex directly to rule-sensitive neurons in premotor cortex. After much practice, the direct network is sufficient to control the behavior, without prefrontal involvement. The model is implemented as a biologically-detailed neural network constructed from spiking neurons and displaying a biologically plausible form of Hebbian learning. The model successfully accounts for single-unit recordings and human behavioral data that are problematic for other models of automaticity.The dissertation also presents the results from two experiments investigating the nature of what is automatized after lengthy practice with a rule-guided behavior. The results of both experiments suggest that an abstract rule, if interpreted as a verbal-based strategy, was not automatized during training, but rather the automatization linked a set of stimuli with similar values on one visual dimension to a common motor response. The experiments were designed to test the Cortex Automatizes Rules Model. The present results support this model and suggest that the projections from visual cortex to prefrontal and premotor cortex are restricted to visual representations of the relevant stimulus dimension only.

This project proposes the use of technology to tackle the issues being faced by pet owners. Pet owners who are working individuals and live alone face a huge problem, they are not able to feed their pets on time. The method approached to solve this issue was to construct an IoT-based Automatic Pet feeder, it is one of the new technologies used for feeding pets and maintaining the diet of their pets by feeding them on time. It is built with Raspberry Pi 3B+ as its core. The automatic pet feeder will automatically dispense a predetermined amount of food and water to the bowl as per the settings made by the owner. The feeder assures hygiene and quality of the food. It will solve a huge problem that pet owners face i.e., making sure that each pet has access to a healthy amount of food throughout the day, regardless of the owner’s schedule, by doing so the pet’s health will be well maintained.

Treball desenvolupat en el marc del programa "European Project Semester"., ”Automated assembly lines are a fundamental part of today’s manufacturing industry” [1]. The future of manufacturing is near, and most manufacturers are looking towards industry 4.0, while AkzoNobel Vilafranca partly is using manual tools to manufacture their own products. AkzoNobel has reached its limits when it comes to its production line, which means that the company must make some necessary and significant changes to reach their goals about increasing its capacity. Production automation focuses on reducing the physical effort of workers to focus on quality control to achieve the high standards of products sought by the company. Automation can be the necessary change for a better outcome in the end. By using automated multi-stations instead of human resources, you can minimize labour costs and reach a much higher production output [2]. While minimizing the costs and need for human resources, there are a lot of opportunities in hiring people for future development in the company. To compete with your opponents on the field, you must be willing to make the changes for the future of the company. There are two main aims of this project. The first one is to propose a fully automatic filling process for AkzoNobel Vilafranca. The second aim is to develop a business case study that would show the feasibility of the proposal in a real-life situation. Such a process would allow AkzoNobel management to move their human resources to product preparation, without the need for manual labour in filling processes. As a result, AkzoNobel would be able to increase its production capacity and possibly expand to other markets. This report aims to present the work of the AkzoNobel team over the first part of the European Project Semester (EPS), on the project of automatic filling lines. Firstly, the report begins with an introduction to the EPS program, the project team, supervisors, and the company. Secondly, the scope of the project is defined, giving details of the current situati, Incoming

PurposeTo assess the generalizability of a deep learning-based algorithm to segment the ellipsoid zone (EZ).MethodsThe dataset consisted of 127 spectral-domain optical coherence tomography volumes from eyes of participants with USH2A-related retinal degeneration enrolled in the RUSH2A clinical trial (NCT03146078). The EZ was segmented manually by trained readers and automatically by deep OCT atrophy detection, a deep learning-based algorithm originally developed for macular telangiectasia Type 2. Performance was evaluated using the Dice similarity coefficient between the segmentations, and the absolute difference and Pearson's correlation of measurements of interest obtained from the segmentations.ResultsWith deep OCT atrophy detection, the average (mean ± SD, median) Dice similarity coefficient was 0.79 ± 0.27, 0.90. The average absolute difference in total EZ area was 0.62 ± 1.41, 0.22 mm2 with a correlation of 0.97. The average absolute difference in the maximum EZ length was 222 ± 288, 126 µm with a correlation of 0.97.ConclusionDeep OCT atrophy detection segmented EZ in USH2A-related retinal degeneration with good performance. The algorithm is potentially generalizable to other diseases and other biomarkers of interest as well, which is an important aspect of clinical applicability.

Els models de Deep Learning (DL) s'estan convertint en un dels camps del Machine Learning (ML) més extesos en el seu ús. Tot i això, la gran quantitat de paràmetres i operacions que realitzen les xarxes neuronals les fan massa complexes per a que un humà pugui entendre les accions i decisions que prenen. L'opacitat del funcionament d'aquests models requereix d'algun sistema que expliqui i justifiqui les prediccions que se n'obtenen. Aquesta justificació és necessària en molts casos del món real, des del camp de la medicina al legislatiu, on es tracta informació sensible que requereix una confiança en que les decisions del model siguin correctes. En el cas d'aquest projecte, la motivació de les explicacions ve donada per l'aplicacació del DL al camp de les companyies asseguradores, entre les quals hi ha el Grup Catalana Occident (GCO) que és on s'ha desenvolupat aquest projecte. L'explicabilitat en models d'aprenentatge automàtic, i concretament de DL ha evolucionat molt al llarg dels darrers anys, i en l'actualitat hi ha una gran quantitat de mètodes per a obtenir explicacions. L'objectiu d'aquest projecte és la recopilació i homogeneització de diversos algoritmes d'explicabilitat enfocats al Deep Learning, principalment agrupats en mètodes Globals (explicacions generals del model) i mètodes Locals (explicacions de prediccions concretes). Amb les definicions i la unificació de notació d'aquests mètodes s'ha fet una implementació bàsica de tots ells sobre el conjunt de dades públic ''Adult Data Set'', i una posterior adaptació per un dataset privat de l'empresa, que serveixi com a guia per a la futura implementació d'una llibreria que permeti aplicar out-of-the-box aquests algoritmes en nous projectes que utilitzin models de DL., Deep Learning (DL) models are becoming one of the most widespread fields of Machine Learning (ML) in use. However, due to the large number of parameters and operations performed by neural networks, they make them too complex for a human to understand the actions and decisions they make. The opacity of the networks decision making requires some system that explains and justifies the predictions obtained. This justification is necessary in many cases in the real world, from the field of medicine to legislature, where sensitive information is involved that requires user trust in the model's decision making. In the case of this project, the motivation for the explanations is given by the application of DL in the field of insurance companies, among which is the Grup Catalana Occidente (GCO) which is where this project has been developed. The explainability in machine learning models, and specifically in DL, has evolved a lot over the last few years, and today there are a lot of methods to get explanations. The aim of this project is to compile and unify various explainability algorithms focused on deep learning, mainly grouped into Global methods (general explanations of the model) and Local methods (explanations of specific predictions). Given the definitions and unification of these methods, a basic implementation of all of them has been made on a public data set ''Adult Data Set'', and a subsequent adaptation on a private dataset that belongs to the company, which will serve as a guide for future implementation of explainable DL algorithms for projects that are built with neural networks.

Perkembangan teknologi informasi semakin maju dan cepat. Banyak teknologi masuk ke bidang lain seperti pertanian. Dengan berkembangnya teknologi, sektor pertanian juga mengalami perkembangan pola bercocok tanam tanpa media tanah yang dikenal dengan hidroponik. Hidroponik membutuhkan pengairan melalui pipa yang dialiri pompa listrik yang dioperasikan secara manual. Perlu ada tenaga manusia yang standby untuk mengoperasikannya. Ada kemungkinan manusia bisa lalai dan menyebabkan sistem irigasi terganggu sehingga tanaman mati. Membuat pompa berjalan secara otomatis adalah salah satu cara untuk mengurangi terjadinya masalah tersebut. Sistem pompa otomatis ini akan dikendalikan oleh mikrokontroler untuk mengairi tanaman secara otomatis dan dapat memenuhi sumber hara yang dibutuhkan. Salah satu penerapan sistem pengendalian irigasi adalah perlunya menjaga air dan unsur hara sesuai dengan ukuran yang tepat untuk tanaman tertentu. Dengan sistem irigasi otomatis pada tanaman hidroponik dapat menghemat energi dan mempermudah dalam bekerja dalam merawat tanaman dengan media tanam hidroponik. Salah satu teknologi yang digunakan adalah aplikasi Arduino IDE. Untuk terhubung dengan alat sensor dan aplikasi website, diperlukan mikrokontroler sebagai pusat kendali. Pembuatan aplikasi website menggunakan bahasa pemrograman PHP. Sedangkan untuk mikrokontroler menggunakan bahasa pemrograman C++, The development of information technology is increasingly advanced and fast. Many technologies enter into other fields such as agriculture. With the development of technology, the agricultural sector has also experienced developments in the pattern of farming without soil media, known as hydroponics. Hydroponics requires irrigation through pipes that are flowed by an electric pump that is operated manually. There needs to be a standby human power to operate it. It is possible that humans can be negligent and cause the irrigation system to be disrupted so that the plants die. Making the pump run automatically is a way to reduce the occurrence of these problems. This automatic pump system will be controlled by a microcontroller to irrigate the plants automatically and can meet the required nutrient sources. One of the implementations of an irrigation control system is that it is necessary to keep water and nutrients in accordance with the right size for certain plants. With an automatic irrigation system on hydroponic plants, it can save energy and make it easier to work in caring for plants with hydroponic growing media. One of the technologies used is the Arduino IDE application. To connect with sensor tools and website applications, a microcontroller is needed as a control center. Making website applications using the PHP programming language. As for the microcontroller using the C + + programming language.

Aquest Treball de Fi de Grau té per objectiu el tractament de les dades obtingudes en el procés de reparació de les avaries que afecten la flota d'autobusos d'una empresa de transport urbà de viatgers, amb la finalitat de crear un sistema recomandador d'actuacions enfocades en el manteniment predictiu. El procés seguit ha estat analitzar aquestes dades i testejar diferents models predictius, per tal de trobar la manera més encertada possible de predir quan està previst que determinat sistema d'un vehicle pateixi una avaria i quin serà el motiu que la produirà. Per tal d'aconseguir-ho, les dades subministrades per l'empresa s'han complementat a partir dels manuals de les diferents marques de vehicles, s'han processat i s'han transformat per tal de millorar l'encert, alhora, s'ha treballat amb diferents models predictius per tal de trobar aquells que millors resultats donen. Finalment, el que s'ha obtingut és un sistema que recomana accions de treball sobre aquells sistemes dels vehicles que previsiblement s'avariaran abans i produiran una irregularitat major en el servei., This Final Degree Thesis aims to process the data obtained in the process of repairing breakdowns that affect the bus fleet of an urban travel transport company, in order to create a recommender system focused on predictive maintenance actions. The process followed has consisted of analyzing this data and testing different predictive models, to find the most accurate way possible to predict when a certain system of a vehicle is expected to suffer a breakdown and what will be the reason for it. In order to achieve this, the data supplied by the company has been supplemented with the manuals of the different vehicle brands, processed and transformed to improve success, while working with different predictive models to find those that give better results. Finally, what is obtained is a system that recommends work actions on those vehicle systems that will foreseeably break down earlier and will produce a greater irregularity in the service.

This work presents an assessment of the potential of model predictive control (MPC) of a Dutch polder system. The system drains to the Linge river and includes 13 weirs, 4 hydraulic gates and 4 large pumping stations each equipped with multiple pumps, managed by the Water Board Rivierenland. The management of the system must comply with several goals: keep the water levels within the bounds of safety, pump out the excess water at minimum cost or CO2 emission, but always have enough water for irrigation and shipping. To achieve these goals there are weirs regulating the water level in different pools, pumping stations to pump water in and out and gates to let water in and out by free flow when possible. These pumping stations consume large amounts of energy. We propose multi-objective mixed-integer optimization by using goal programming to prioritize different operational objectives. For the control of the pumps mixed-integer optimization is used, which makes it possible to not only model the energy consumption of the pumps while in operation, but also to model if the pumps are turned on or off. The control system is implemented using RTC-Tools, an open-source software tool to implement MPC. It is demonstrated that the proposed control system implementation can comply with the operational goals of the water board: keeping the water levels within the bounds while reducing the operational costs. The proposed control system has been tested numerically on data from the year 2013, and it is shown that it highly outperforms the current operation.

[ES] El objetivo de este proyecto es diseñar un controlador de movimiento bio-inspirado capaz de adaptarse a los cambios en la dinámica de formaciones de robots submarinos modulares que pueden alterar su distribución física. Los robots usados pertenecen al laboratorio REMORA de la Universidad Técnica de Dinamarca (DTU). Estos han sido diseñados para llevar a cabo tareas de inspección, reparación y mantenimiento de forma autónoma en entornos submarinos y pueden unirse físicamente entre sí en formaciones de dos o mas unidades para compartir recursos y funciones, o como medida de mitigación de fallos. Estas formaciones dan lugar a sistemas con dinámicas distintas no modeladas que el controlador debe conocer para poder proporcionar un control de movimiento con mayores prestaciones. Para lograr esto, la estructura de control diseñada introducirá técnicas basadas en neurociencia (específicamente en modelos computacionales del cerebelo) y aprendizaje automático, implementándolas en un módulo de aprendizaje que actualiza en tiempo real un modelo de la dinámica del sistema. De esta manera, los robots modulares pueden continuar funcionando de manera fiable y con un mayor nivel de desempeño sin la necesidad de rediseñar manualmente el modelo físico y la estructura de control necesarios para cada formación posible. Este controlador se comparará con otras alternativas tradicionales para evaluar su rendimiento y validez. En el desarrollo del proyecto se realizará inicialmente una revisión de la literatura en el campo de controladores de aprendizaje y de modelos computacionales del cerebelo para escoger aquella alternativa que mejor se adapte al problema formulado. En la implementación se recurrirá en primer lugar a simulaciones físicas del sistema para iterar rápidamente en el diseño del controlador y a continuación la solución adoptada se trasladará a los sistemas robóticos para validar su funcionamiento en distintos escenarios. Se estudiará la aplicación a tareas típicas de gui, [EN] New challenges in inspection, maintenance and repair tasks in offshore installations have given rise to an increasing demand of small autonomous and cooperative underwater vehicles that provide the required versatility to face a wide array of complex missions. An envisioned solution is to use multiple modular vehicles that can attach to each other rigidly into formations to share resources and extend their carrying or tool usage capabilities. These vehicle formations may be required to execute complex maneuvers with high accuracy to interact with the underwater installations in a precise manner. These demanding motion capabilities can be framed as a trajectory tracking problem, which requires knowledge of the inverse dynamics of the vehicles in order to anticipate the forces to be executed by their motors to set the vehicles into a desired kinematic state. Since the dynamics model of the vehicle formations is not known, the proposed solution considers a biologically-inspired learning controller that can fit a relevant inverse dynamics model in a short period of time. This controller is based on the ideas presented in [30], and it is shown that it outperforms a classical PI controller with high gains in a set of designed test scenarios, both in simulation and with a real underwater vehicle. For the simulation, the model of a single vehicle and the model of two vehicles based on [23] were implemented to rapidly iterate in the design process and tuning of the control algorithm, as well as to provide part of the validation of the controller capabilities. For the real vehicle validation, the control algorithm was implemented in ROS for both a single vehicle and two vehicles, although only the single-vehicle controller could be tested due to limitations out of the scope of this project.

Internet of Things (IoT) based smart farming is used to automate farming. It is used to monitor the level of environmental condition is the main key factor to improve the yield of efficient crops. The IoT system has a full-duplex communication link based on the internet interface that allows data irrigation and fertilizer schedule to be programmed through an android application (“BLYNK App”). It is used to automate the irrigation and fertilizer supply. Blynk application is used to control the hardware remotely and display the sensor values. ESP8266 is a Wi-Fi module controlled by the blynk application. Irrigation and fertilizer supply can control the two different types. One is automation through IoT and another one is manual. In automation, the whole system is controlled by blynk application through the phone. In manual type, humans can on/off the system. In this way, we also measure the field (or) farm temperature and humidity using IoT we can control anything everywhere. This method is used to improve crop yield. The proposed method is used to reduce manpower, time efficiency, and high cost.

The paper deals with manufacturing or bending of sheet metal by using power operated sheet bending machine. Especiallydiscussion made the productivity analysis of manually or power operated sheet bending machine. Considering manual operation is replaced by power operated devices.It also gives information about limitation of manually operated sheet bending machine and power operated sheet bending machine.

Automated mouse phenotyping through the high-throughput analysis of home cage behavior has brought hope of a more effective and efficient method for testing rodent models of diseases. Advanced video analysis software is able to derive behavioral sequence data sets from multiple-day recordings. However, no dedicated mechanisms exist for sharing or analyzing these types of data. In this article, we present a free, open-source software actionable through a web browser (an R Shiny application), which performs an analysis of home cage behavioral sequence data, which is designed to spot differences in circadian activity while preventing p-hacking. The software aligns time-series data to the light/dark cycle, and then uses different time windows to produce up to 162 behavior variables per animal. A principal component analysis strategy detected differences between groups. The behavior activity is represented graphically for further explorative analysis. A machine-learning approach was implemented, but it proved ineffective at separating the experimental groups. The software requires spreadsheets that provide information about the experiment (i.e., metadata), thus promoting a data management strategy that leads to FAIR data production. This encourages the publication of some metadata even when the data are kept private. We tested our software by comparing the behavior of female mice in videos recorded twice at 3 and 7 months in a home cage monitoring system. This study demonstrated that combining data management with data analysis leads to a more efficient and effective research process., Peer Reviewed

Objectives: To provide dental practitioners and researchers with a comprehensive and transparent evidence-based overview of the characteristics of literature regarding initiatives of robot technology in dentistry. Data: All articles in which robot technology in dentistry is described, except for non-scientific articles and articles containing secondary data (reviews). Amongst others, the following data were extracted: type of study, level of technological readiness, authors’ professional background and the subject of interaction with the robot. Sources: Bibliographic databases PubMed, Embase, and Scopus were surveyed. A reference search was conducted. The search timeline was between January 1985 and October 2020. Study selection: A total of 911 articles were screened on title and abstract of which 161 deemed eligible for inclusion. Another 71 articles were excluded mainly because of unavailability of full texts or the sole use of secondary data (reviews). Four articles were included after hand searching the reference lists. In total, 94 articles were included for analysis. Conclusions: Since 2013 an average of six articles per year concern robot initiatives in dentistry, mostly originating from East Asia (57%). The vast majority of research was categorized as either basic theoretical or basic applied research (80%). Technology readiness levels did not reach higher than three (proof of concept) in 55% of all articles. In 84%, the first author of the included articles had a technical background and in 36%, none of the authors had a dental or medical background. The overall quality of literature, especially in terms of clinical validation, should be considered as low., Learning & Autonomous Control

Objectives: To provide dental practitioners and researchers with a comprehensive and transparent evidence-based overview of physical robot initiatives in all fields of dentistry. Data: Articles published since 1985 concerning primary data on physical robot technology in dentistry were selected. Characteristics of the papers were extracted such as the respective field of dentistry, year of publication as well as a description of its usage. Sources: Bibliographic databases PubMed, Embase, and Scopus were searched. A hand search through reference lists of all included articles was performed. Study selection: The search timeline was between January 1985 and October 2020. All types of scientific literature in all languages were included concerning fields of dentistry ranging from student training to implantology. Robot technology solely for the purpose of research and maxillofacial surgery were excluded. In total, 94 articles were included in this systematic review. Conclusions: This study provides a systematic overview of initiatives using robot technology in dentistry since its very beginning. While there were many interesting robot initiatives reported, the overall quality of the literature, in terms of clinical validation, is low. Scientific evidence regarding the benefits, results and cost-efficiency of commercially available robotic solutions in dentistry is lacking. The rise in availability of open source control systems, compliant robot systems and the design of dentistry-specific robot technology might facilitate the process of technological development in the near future. The authors are confident that robotics will provide useful solutions in the future but, strongly, encourage an evidence-based approach when adapting to new (robot) technology., Learning & Autonomous Control

Since the 13th of October 2017, the Tropospheric Monitoring Instrument (TROPOMI) aboard ESA’s Sentinel 5-Precursor (S5-P) satellite enables daily global measurements of carbon monoxide (CO) total column con- centrations at an unprecedented spatial resolution of 7×5.6 km. TROPOMI has the ability to detect distinct pollution plumes, arising from point source emissions, from which emission rates can be derived. We in- vestigate the potential of CO column concentrations observed by TROPOMI to estimate the CO emissions of point sources on an operational level. This study developed a Python framework that for pre-defined point sources automatically detects pollution plumes and from which it estimates CO emissions using a mass bal- ance approach directly from single overpass CO observations. The algorithm is based on concepts from the computer vision to identify the plume and extract the plume center line while respecting the plume orienta- tion. The emission rate is approximated from flux profiles through multiple plume cross-sections following the plume center line. The performance of the developed framework and its potential is demonstrated by the application on 132 identified steel plant facilities over a time period of more than 2.5 years. Currently the lack of accessible and quality-wise good data limits spatial or even temporal comparison of CO emissions from steel plants. Therefore the control and understanding of emission rates could greatly benefit from the proposed approach. In total we obtained 1,774 emission estimates for 97 facilities. Up to 119 measurements per facility are derived where for the majority of the facilities the average number of measurements is around 10. The obtained time series showed large variation in the distribution of measurements over time as well as the emission values itself. For a number of higher emission values, that exceeded up to 2 times the aver- age emission, measured for e.g. the Bhilai Steel Plant, India, the outliers corres, Civil Engineering | Environmental Engineering

I detta projekt behandlas en problembeskrivning med syfte att designa ett produktförslag till ett automatiskt roterande tårtfat. Produkten ska uppfylla de behov och funktioner som är önskvärda av kunder. Målet är att förenkla tårtskapares arbete genom att designa produktförslaget. Tårtfat används från uppbyggnad av tårtan fram till serveringen och är det viktigaste verktyget vid skapandet av tårtor. Under arbetets gång har gruppen utvecklat nya kunskaper inom innovation och design och har dessutom stärkt rollen som ingenjör. Genom projektets gång har arbetet haft ett fokus på produktutvecklingsprocessen där olika undersökningar och tester genomförts. Arbetet har inriktat sig mot designprocessen med fördjupning i analyser och flera iterationer av utformningen. Rapporten presenterar projektets produktutvecklingsprocess och den idé som tagits fram för att produkten ska fungera. Rapporten innehåller även olika teorikapitel med fakta om elmotorer, material och CE-märkning. Resultatet av projektet blev ett designförslag för det automatiska tårtfatet med beskrivning och sprängskiss av hur komponenter ska kopplas samman. Vidare presenteras ett designförslag för en extra bytbar överdel som används vid servering. Utseendet presenteras med hjälp av datorbaserade hjälpmedel och de mekaniska aspekterna diskuteras och finns som förslag för vidareutveckling av produkten., This project deals with a problem description with the aim of designing a product proposal for an automatic rotating cake stand. The product will meet the needs and functions that are desired by the customers. The goal is to simplify the work for a cake maker by designing the product. Cake stands are used from construction of the cake until serving and are the most important tool in cake making. During the work, the group has developed new knowledge in innovation and design and has also strengthened their roles as engineers. Throughout the project, the work has centered on the product development process where various investigations and tests have been carried out. The work has focused on the design process with in-depth analysis and several iterations of the design. The report presents the project's product development process and the idea developed for the product. The report also contains a number of different theoretical chapters with facts about electric motors, materials and CE marking. The result of the project was a design proposal for the automatic cake stand with a description and exploded view of how components should be connected. Furthermore, a design proposal is presented for an extra exchangeable top that is used when serving. The appearance is presented with the help of CAD while the mechanical aspects are discussed and is available as a proposal for further development of the product.

This project is a master’s thesis in collaboration with Nomor, a company that offers pest control services. They had the need for the development of an automated mechanical rat trap for sewage systems. Brown rats populate the sewage systems in many cities, as it offers shelter and abundant food resources. Rats have been observed carrying various zoonotic diseases which are a health risk to people and animals, as rats may enter households via the sewage system. This report details the design and development process for a prototype for a striking mechanism, which will act as the foundation for a future trap. The process follows the theory described by Ulrich and Eppinger. The project resulted in the generation of 8 different concepts, which were screened down to a final pneumatic concept. The selected concept was then developed into a physical prototype, able to detect heat sources. A Simulink model was developed to support the detail design. The model calculates resulting impact energy from various input parameters, such as working pressure and moving mass. The prototype functions up to 3 bar with an impact energy of approximately 6 Joules.

Scientific computation problems have been faced with the need to analyze increasing amounts of data as part of their application workflows, and the science-based model is being combined with big data and machine learning models to solve complex problems and phenomena [1][2]. The machine learning workflow is composed of some reproducible steps that can be executed as a pipeline to build a model efficiently by saving iteration time, helping in debugging and detecting [3]. Currently, businesses and researchers are investigating and improving the methodology of developing and deploying machine learning workflows in both training and inference phases, which helps the data science team focus on their requirements and the data engineer team deploy and operate machine learning workflows efficiently and automatically [4]. This work presents an architecture for automatic machine learning workflows, which provides capabilities of monitoring and automatic management on the end-to-end life-cycle of machine learning workflows, including tracking and observing at the training stage, and releasing, monitoring, deployment, auto-detecting and infrastructure management at the inference stage. To validate feasibility, we have conducted a case study based on our architecture and deployed it in the cloud, and showed its automation.

Despite of having many rules and regulations in for transportation traffic, it is still to this day a chaotic environment sometimes. Therefore, finding traffic solution is a work of art in civil engineering world. However, the technology of internet and navigation solution is evolve every day, and this opens many opportunities to increase the quality of current traffic status or even solve problems that people do not even know it existing. It is such a fascinating time we are living where electronics components embedded with latest technology are affordable for hobbyist, and it means more beautiful mind can involve in finding delicate solutions, and current manual toll station is one of the problems that need a better solution. GPS technology and Google database now reach a level that is precisely, smart and fast enough on a portable smartphone to change the traditional toll to a new form by removing all of the obstacle throughout fee paying process at these station and to make it environmental friendly. This project is one of the solution by realizing an on the go payment virtual toll station, replace all the work of manual toll to a smarter automatic one.