Vilnius Tech DSpace Repository
Not a member yet
52562 research outputs found
Sort by
2024 International Conference "Electrical, Electronic and Information Sciences“ (eStream)
No full textPredicting Time Complexity of TensorFlow Lite Models
No full textThe Internet of Things (IoT) revolutionizes global communication and data exchange, while edge computing optimizes data transmission and enhances security, adaptability, and cost-effectiveness. However, integrating computer vision into IoT systems faces challenges in adapting algorithms to limited computational power and memory resources. This paper investigates the performance of deep vision models designed for low-power systems, focusing on inference time. Through a comprehensive experiment, various structures of convolutional neural network (CNN) models are evaluated based on their layer configurations and inference time. Systematically analyzing these components, the research establishes a predictive formula for inference time estimation based on the model architecture. The results reveal dependencies between CNN layer complexity and inference efficiency, guiding optimal configurations for edge device deployment. This analysis offers insights for designing efficient deep vision models tailored for low-power systems.Taip / Ye
Žaliųjų finansų modelis tvariai miestų be atliekų plėtrai
No full textGreen finance is a way of financing climate change mitigation, adaptation and other environmental challenges. Along with the green economy concept, green finance contributes to disseminating global environmental challenges and how these challenges can be overcome. One area of environmental challenge is waste and sustainable waste management for sustainable urban development, where the green economy and green finance disciplines can have a significant positive impact.
The dissertation addresses the insufficiently effective use of the green finance model applicable to reducing municipal waste and following the “zero-waste” principles in the scientific literature and practice. It aims to develop and empirically test a green finance model that would lead to rational economic decisions for the sustainable development of zero-waste cities.
The dissertation consists of an introduction, three chapters, and general conclusions. The First Chapter aims to determine the green finance concept within related economic theories, concepts, methods, and models. It provides insights into other interrelated concepts, such as bio-economy, low-carbon finance, climate finance, and socio-environmental finance. Additionally, it provides a review of the green finance instruments. The Second Chapter investigates the role of waste management within economic theories and establishes the concept of zero-waste and inter-municipal cooperation. It also shapes the strategy for developing zero-waste cities with the active involvement of green finance mechanisms. The Second Chapter describes the methods for forming a green finance model for developing zero-waste cities. The Third Chapter provides empirical approval results of the green finance model for developing zero-waste cities on the Lithuanian regional level.
A set of general conclusions and recommendations for future research complete the dissertation. Five publications on the discussed study were published: two in the reviewed scientific journals, and three in conference proceedings.Žalieji finansai – tai finansavimo būdas, kuriuo siekiama sušvelninti klimato kaitą, prisitaikyti prie jos ir spręsti kitas aplinkosaugos problemas. Kartu su žaliosios ekonomikos koncepcija, žalieji finansai prisideda prie pasaulinių aplinkosaugos iššūkių ir būdų, kaip šiuos iššūkius galima įveikti, sklaidos. Viena iš aplinkosaugos problemų sričių – atliekos ir tvarus atliekų tvarkymas siekiant tvarios miestų plėtros, kur tiek žaliosios ekonomikos, tiek žaliųjų finansų disciplinos gali turėti didelį teigiamą poveikį.
Disertacijoje nagrinėjama problema, kad mokslinėje literatūroje ir praktikoje nepakankamai efektyviai taikomas žaliųjų finansų modelis, siekiant mažinti komunalines atliekas ir laikytis nulinių atliekų principų. Disertacijos tikslas – sukurti ir empiriškai patikrinti žaliųjų finansų modelį, kurio taikymas leistų priimti racionalius ekonominius sprendimus siekiant darnios miestų be atliekų plėtros.
Disertaciją sudaro įvadas, trys skyriai, bendrosios išvados. 1 skyriuje siekiama apibrėžti žaliųjų finansų sąvoką susijusiose ekonomikos teorijose, koncepcijose, metoduose ir modeliuose, pateikti įžvalgas apie kitas tarpusavyje susijusias sąvokas, tokias kaip bioekonomika, mažo anglies dioksido kiekio technologijų finansavimas, klimato kaitos finansavimas ir socialinės aplinkosaugos finansavimas, taip pat apžvelgti žaliųjų finansų produktus. 2 skyriuje nagrinėjamas atliekų tvarkymo vaidmuo ekonominėse teorijose, nustatoma „zero-waste“ sąvoka ir savivaldybių bendradarbiavimas. Šiame skyriuje taip pat formuojama miestų, kuriuose nėra atliekų, plėtros strategija, aktyviai dalyvaujant žaliųjų finansų mechanizmams. 2 skyriuje aprašomi metodai, taikomi formuojant žaliųjų finansų modelį, skirtą miestų be atliekų plėtrai. 3 skyriuje pateikiami žaliųjų finansų modelio, skirto nulinių atliekų plėtrai, empirinio patvirtinimo Lietuvos regioniniu lygmeniu rezultatai.
Disertaciją užbaigia bendrosios išvados. Aptariamo tyrimo tema paskelbtos 5 publikacijos: 2 – recenzuojamuose mokslo žurnaluose, 3 – konferencijų
medžiagoje.Taip / Ye
Improvement of control process for hot recycling in asphalt mixing plant of transport infrastructure bituminous pavement
No full textDisertacijos darbo aktualumas yra susijęs su kartotiniu senos bituminės kelio dangos panaudojimu gaminant naujus asfalto mišinius. Tikslas yra pateikti mokslinį įvertinimą dėl senos bituminės kelio dangos remonto metu susidarančių asfalto granulių pakartotinio panaudojimo, jas tinkamai paruošiant pridėjimui į naujų asfalto mišinių gamybą bei šiuos mišinius gaminant. Rengiant darbą buvo kompleksiškai ištirtas senos kelio bituminės dangos pakartotinio naudojimo ūkis Lietuvoje, susidarančių pakartotinai naudojamų dangų technologiniai paruošimo ir pridėjimo asfalto gamybos metu būdai ir principai. Kiekybiškai įvertinta technologinių procesų įtaka gaminamo mišinio savybių kaitai. Šiam tikslui pasiekti išanalizuoti mokslo darbai, atliktas skaitinis modeliavimas ir eksperimentiniai tyrimai.
Disertaciją sudaro įvadas, trys skyriai, bendrosios išvados, naudotos literatūros ir autoriaus publikacijų disertacijos tema sąrašai, santrauka anglų kalba ir vienas priedas.
Įvadiniame skyriuje pateikta tiriamoji problema, darbo aktualumas ir tyrimų objektas. Formuluojamas darbo tikslas ir uždaviniai, aprašoma tyrimų metodika, darbo mokslinis naujumas, darbo rezultatų praktinė reikšmė, ginamieji teiginiai. Įvado pabaigoje pristatomos disertacijos tema autoriaus paskelbtos publikacijos.
Pirmajame skyriuje išanalizuoti mokslo darbai, susieti su senos asfalto dangos pakartotiniu panaudojimu. Aptariami naudojami būdai ir priemonės, skirti padidinti naudotos asfalto dangos pakartotinį naudojimą ir šio naudojimo ekonominis efektas. Nagrinėjama senos asfalto dangos pakartotinio naudojimo strategija.
Antrajame skyriuje pateikiama ekspertų nuomonių suderinamumas ir išgryninamas šalyje asfalto dangos regeneravimo santykinį kiekį didinančių veiksnių reikšmingumas, norminių dokumentų analizė, strategijos grėsmių ir iš jų kylančių iššūkių analizė. Išanalizuoti ir susisteminti senos asfalto dangos pridėjimo skirtinguose gamybos etapuose asfalto gamykloje būdai. Pateiktas karštojo regeneravimo metu vykstančių vandens garinimo matematinis modelis.
Trečiajame skyriuje pateikiami eksperimentinių tyrimų rezultatai, aptarti šiltnamio efektą sukeliančių dujų emisijų mažinimo būdai. Nagrinėjami ir vertinami asfalto dangos asfalto maišytuve technologijos parametrai ir jų pagerinimo susistemintieji būdai.
Disertacijos tema paskelbti šeši straipsniai: trys – mokslo žurnaluose, įtrauktuose į Clarivate Analytics Web of Science duomenų bazę ir turinčiame citavimo rodiklį, du – tarptautinių konferencijų leidiniuose, referuojamuose kitų tarptautinių duomenų bazių leidiniuose (Elsiever, Springer), vienas – tarptautinės konferencijos referuojame leidinyje.The relevance of the dissertation is related to the reuse of old bituminous pavement in the production of new asphalt mixtures. The aim is to provide a scientific assessment of the re-use of asphalt grains formed during the repair of an old bituminous road surface, to prepare them adequately for addition to the production of new asphalt mixtures and to produce these mixtures. The energy requirements were determined for adding used asphalt pavement to produce a new asphalt mix in the plant under the influence of moisture at different amounts of the added material. Scientific works were analysed, and numerical modelling and experimental research were performed to achieve this goal.
The dissertation consists of an introduction, three chapters, general conclusions, lists of used literature and the author’s publications on the dissertation topic, a summary in English and one appendix.
The introductory chapter presents the research problem, the work’s relevance, and the research object. It formulates the work’s aim and tasks and describes the research methodology, the scientific novelty of the work, the practical significance of the work results, and the defence statements. The introduction closes with publications by the author on the topic of the dissertation.
The first chapter analyses the research on reusing old asphalt pavement, starting but not limited to preparation processes (milling, crushing, sieving, and storage) and the impact of preparation on technological production processes. It discusses methods and measures to increase the reuse of used asphalt pavement and their economic impact. Also, the chapter considers the strategy of reusing old asphalt pavement.
The second chapter presents expert opinions and the significance of factors increasing the country’s relative amount of asphalt pavement recycling. It provides analyses of regulatory documents and strategy, threats and the challenges arising from them. Furthermore, it presents a mathematical model of the excess water evaporation during hot recycling.
The third chapter presents the experimental studies’ results, and methods of reducing greenhouse gas emissions are discussed. It analyses and evaluates the parameters of the asphalt pavement in the asphalt mixer technology and the systematic ways of their improvement.
Six articles were published on the dissertation’s topic: three in a scientific journal included in the Clarivate Analytics Web of Science database with a citation rate, two in publications of international conferences, referenced in other international databases (Elsevier and Springer), and one in international conference reference paper.Taip / Ye
Comparison of Object Detection Algorithms for the Task of Person Detection on Jetson TX2 NX Platform
No full textThis article is an investigation for the best suited algorithm for object detection on Jetson TX2 NX platform. Previous research aimed at determination of the quickest and the most accurate algorithm for obj ect detection and tracking of vehicles. Current research uses previous article results and tries to apply found models on Jetson TX2 NX platform. The general goal is detection of a person on crosswalk. The model should have higher speed, than better accuracy in this case, because detection speed is a crucial point. This research includes applying of the YOLOv8m with 18 classes, which may appear on the road, YOLOv8s with only person class, DetectNet_v2 with ResNetl0 as a backbone and smaller version of this model. Mentioned models will be compared using trtexec tool on Jetson TX2 NX.Taip / Ye
Kernel Filter-Based Image Construction from Scanning Electrochemical Microscopy Approach Curves
No full textA scanning electrochemical microscope (SECM) with artificial intelligence could generate the sample's activity image from approach curves measured at several points, minimizing the time of measurement and calculating sample activity in points of interest. The time used to shape the separation between the feature space and regression line in the convolutional process for CNNs and DNNs constitutes a significant contribution to the accuracy of the AI model performance. Kernel functions and pre- processing of synthetic data can achieve higher efficiency and localization precision by applying them to the initial layers of an MLP network; assuming an infinite-width network, we can use the theory from NTK to consider the kernel shape and function. In this paper, we compare model accuracy and performance with and without rescaled data for under-sampled data sets and determine the effectiveness of CNN Fourier kernel interpolation filtering to detect high-frequency features for non-image data in multi-layer perceptron (MLP) networks. The results show how engineered feature mapping and data shaping can affect the convergence of several types of Gaussian and Fourier mapping methods for training and validation convergence with custom activation and modified filter functions, as well as model learning of highly sampled features in proprietary data bands.Taip / YesTaipei Tech - Vilnius Tec
Lightweight composite materials made of paper sludge and corn starch /
No full textNovel composite materials have been prepared by primary paper sludge (PS) and biopolymer (modified corn starch (MCS). The MCS is used as a binder. The aim of the research was to investigate the properties, microstructure and durability (water erosion) of composite materials (CM). The CM were formed from 100 wt% of paper sludge and 20 wt% of various concentrations (2.5%, 5.0% and 7.5%) of MCS were added. Corn starch was heated 3 h at 150 °C temperature. The CM produced had a bulk density of 0.89–1.18 g/cm3, compressive strength at 10% relative deformation 4.8–9.3 MPa, thermal conductivity 0.092–0.132 W/m·K. Additionally, this study underscores that the CM show good durability (resistance to water erosion) and hygrothermal qualities. In fact, the MSC solution incorporated in the paper sludge showed improved physical-mechanical properties and durability of the CM. The research shows that CM can be formed from paper sludge and MCS, and these can be used as indoor masonry units because they also have good mechanical, hydroscopic and thermal insulation properties.vol. 2
Energy–saving in hydraulic drives in experimental approach and biomimetric similarity
No full textThe current article presents the research results of the proposed energy-saving way in hydraulic drives based on a biomimetic approach. Experimental measuring for investigation energy consumption by different types of high-pressure hoses and its influence on the inside fluid flow are presented. The methodology presented in this paper for experimental measurement is based on fluid pressure measuring and high-pressure hose vibration measuring analysis. In the research, by experimental measuring, different braided high-pressure hoses, used in modern hydraulic drives, were compared. The research was performed to describe a high-pressure hose behavior dependency on the fluid flow inside and its influence on the energy parameters of the system. The proposed concept for energy-saving in a hydraulic drive is based on biomimetic approaches. In the research it is disclosed that, by using different types of high-pressure hoses, a reduce of power losses and an increasing hydraulic drive efficiency can be achieved, by replacing one type of a high-pressure hose with another type, during machinery hydraulic drive maintenance or on design stage
Exploration of Genetic Algorithm-Driven Hyperparameter Optimization for Convolutional Neural Networks
No full textConvolutional Neural Networks (CNNs) have been widely used in artificial intelligence (AI) research. However, the time-intensive nature of hyperparameter optimization remains a major challenge. In this study, we investigated three distinct CNN architectures, namely, ResNet, Multi-class, and ConvLSTM, by employing Genetic Algorithm (GA). The GA-driven optimization process efficiently adapts the parameters, resulting in significant improvements in validation accuracy and reduced training time. Moreover, our findings indicate that compared to exhaustive manual exploration, GA substantially reduces the time required for hyperparameter optimization, thereby offering a more efficient and effective approach for enhancing model outcomes in reasonable timeframes. This underscores the importance of employing advanced optimization techniques such as GA in AI research to achieve superior model outcomes.Taip / Ye
Natūraliais polimerais modifikuoto molio struktūros ir ilgaamžiškumo tyrimai
No full textThe issue of eco-friendly construction is highly relevant nowadays. Unfired clay materials represent a sustainable alternative to other building products, because of their extremely low carbon footprint. Nevertheless, a pressing need remains to enhance their physical, mechanical, and durability characteristics. To save the great ecological potential of unfired clay materials, the most rational way is to modify them with the help of different biopolymers. Therefore, the dissertation considers the possibilities of different biopolymers (polysaccharides – starch, alginate, and chitosan) to modify clay-based building materials. Different preparation technologies (biopolymers gelation and starch retrogradation processes) were used to evaluate the effect of the examined biopolymers on some properties, structure, and durability of clay materials. Biopolymers and their technological parameters were selected according to their origin, chemical structure, and ability to interact with water. The main principle of clay modification, implemented in this research is the complete replacement of water with a biopolymer solution during the manufacturing process of clay composites. The resulting clay composites with starch or alginate solutions exhibit not only improving mechanical properties but also durability, which could make them very attractive for a wide variety of applications from clay masonry units to plasters. The first chapter presents a comprehensive outlook on contemporary trends in the evolution of sustainable clay-based construction materials. Furthermore, it delves into an exploration of the biopolymers, their distinctive characteristics, and the potential avenues for modification. The second chapter is dedicated to the exploration of the raw materials, processing techniques, testing methodologies and equipment used to evaluate the impact of biopolymers on unfired clay composites. The discussion has also focused on the technology of producing gelated biopolymer aqueous solutions, and the possibilities to technologically modify specific biopolymers using the retrogradation process. Also, this chapter presents the methodology of clay composite testing. The third chapter presents the evaluation of the effect biopolymers have on the physical, mechanical, and hygroscopic properties and durability of clay-based composites, describing interaction mechanisms between clay and biopolymers, and the peculiarities of clay composite structure formation. Nine scientific papers were published on the topic of the dissertation: three papers in scientific journals included in the list of Clarivate Analytics Web of Science database with an impact factor, and four papers in conference proceedings included in the Clarivate Analytics Web of Science and Scopus databases, two – in other international databases. A Lithuanian patent based on the results of the dissertation has been granted, and seven reports were presented at national and international scientific conferences.Taip / Ye