Blekinge Institute of Technology
Not a member yet
13576 research outputs found
Sort by
Governing the commons : code ownership and code-clones in large-scale software development
No full textContext: In software development organizations employing weak or collective ownership, different teams are allowed and expected to autonomously perform changes in various components. This creates diversity both in the knowledge of, and in the responsibility for, individual components. Objective: Our objective is to understand how and why different teams introduce technical debt in the form of code clones as they change different components. Method: We collected data about change size and clone introductions made by ten teams in eight components which was part of a large industrial software system. We then designed a Multi-Level Generalized Linear Model (MLGLM), to illustrate the teams’ differing behavior. Finally, we discussed the results with three development teams, plus line manager and the architect team, evaluating whether the model inferences aligned with what they expected. Responses were recorded and thematically coded. Results: The results show that teams do behave differently in different components, and the feedback from the teams indicates that this method of illustrating team behavior can be useful as a complement to traditional summary statistics of ownership. Conclusions: We find that our model-based approach produces useful visualizations of team introductions of code clones as they change different components. Practitioners stated that the visualizations gave them insights that were useful, and by comparing with an average team, inter-team comparisons can be avoided. Thus, this has the potential to be a useful feedback tool for teams in software development organizations that employ weak or collective ownership. © The Author(s) 2024
NORDEN : Pneumatiskt Gripdon
No full textThis thesis project was carried out in collaboration with Norden Machinery AB, with the goal of developing a pneumatic gripper for handling bottles in a tray packing machine, The project began with extensive ideation through brainstorming sessions and mind mapping, where various gripper concepts were generated. The focus was to develop a solution that is simple, reliable and adaptable to different bottle geometries and sizes. An iterative development process was applied, in which early design steps were repeated alongside systematic evaluations using MET and MECO matrices. Concepts were compared based on cost, strength, durability and flexibility. The chosen solution involves a tool that uses compressed air to inflate a rubber ring, which lifts the bottles by expanding into the cavities created by the arrangement of bottles in the tray. The design was optimized in CAD using FEM analysis and adapted for manufacturability and ease of maintenance. The result demonstrates the potential to streamline bottle handling in automated production lines through a customized yet straightforward pneumatic gripper solution. Detta examenarbete har genomförts i samarbete med Norden Machinery AB med ett syfte att utveckla ett pneumatiskt gripdon för hantering av flaskor i en trågpackningsmaskin. Projektet inleddes med en omfattande idégenerering genom brainstorming och tankekartor där flera alternativa gripdon togs fram. Fokus låg på att komma fram till en lösning som är enkel, driftsäker och flexibel för olika flaskgeometrier och storlekar. En iterativ utvecklingsprocess där steg 1 och 2 upprepades samtidigt som verktyg som MET- och MECO-matriser användes för att systematisk jämföra koncept baserat på kriterier som kostnad, hållfasthet, hållbarhet och flexibilitet. Arbetet resulterade i ett koncept som bygger på användning av ett verktyg där en gummiring blåses upp med tryckluft. Den upplåsta ringen lyfter flaskorna genom att fylla ut de hålrum som uppstår vid flaskornas förpackningskonstellation. Lösningen är optimerad i CAD med hjälp av FEM-analys och är anpassad för tillverkning samt enkel service. Resultatet visar möjligheten att effektivisera flaskhantering inom automatiserade produktionslinjer med hjälp av ett specialanpassat med ändå enkelt gripdon
On the Org of Schema : by Means of Artificial Selection
No full textThis study explores the use of large language models (LLMs) in the selection and generation of Schema.org markup for web pages. The proposed artifact leverages Google Gemini 2.5 Pro to automate the generation of schema markup, which may increase search engine visibility and strengthen search engine optimization (SEO) efforts. The research compares the artifact-generated markup with pre-existing, human-generated schema markup from high-traffic websites in the U.S., evaluating syntactic validity and the schemas’ ability to trigger rich results in Google’s search engine result page. The study finds that while the artifact-generated schemas were more complex and longer than their human-generated counterparts, they exhibited a higher error rate, more warnings, and fewer schema and rich results items, suggesting that they could negatively impact search engine visibility. The analysis also reveals performance characteristics, with the artifact processing an average of 7041 input characters per second at an average processing time of 39 seconds, proving impractical for large-scale application. This work contributes to the emerging field of AI-driven schema generation, highlighting both the potential and the limitations of LLMs in producing high-quality structured data. While the results suggest that LLMs, when curated, could assist in schema generation for smaller-scale applications, further research is needed to address issues of error handling, runtime optimization, and scalability
Minimizing Data Loss in Bluetooth LowEnergy (BLE) : A Parallel Transmission and Application-Layer Modification Approach
No full textBluetooth Low Energy (BLE) is a preferred wireless protocol for Industrial Internet of Things (IIoT) due to its low energy consumption and wide adoption. However, when BLE is accessed via the Web Bluetooth API, it faces challenges such as packet loss, limited control over connection parameters, and restricted throughput. This thesis aims to evaluate the reliability of BLE communication using the Web Bluetooth API and to propose application-layer solutions to reduce packet loss, improve latency, and maintain stable long-term connections in industrial monitoring contexts. An empirical experimental approach was used, involving three sequential experiments that tested different BLE configurations. These included the use of parallel transmission modules, varying data transmission intervals, and timestamp-based packet handling. Performance metrics such as packet loss and jitter were analyzed. Results showed that using two BLE modules in parallel reduced packet loss to zero and minimized jitter. A transmission interval of 2 Hz provided the best trade-off between performance and stability across all packet sizes. Application layer strategies improved connection reliability without requiring hardware level changes. The study shows that BLE communication over the Web Bluetooth API can be stabilized through application-level techniques, enabling real-time and reliable IIoT data transmission with now hardware modification needed
Towards Domain-Grounded Retrieval : Graph Based Prompting for Bug Reports
No full textBackground: Retrieval-Augmented Generation (RAG) systems combine retrieval techniques with language models to produce contextually relevant responses. However, these systems often lack domain-specific grounding and interpretability, which are crucial in fields like software engineering and telecommunications. Objectives: This thesis explores the application of a GraphRAG approach for root cause identification in a software system. It investigates whether integrating a Knowledge Graph (KG) with an LLM can enhance the relevance, accuracy, and interpretability of generated responses. Additionally, it examines whether domain experts find value in combining Large Language Models (LLMs) with an interactive, structured representation of the system under analysis. Methods: The proposed GraphRAG system is compared to a hybrid black-box RAG system, which combines vector-based retrieval with symbolic methods. While hybri-RAG approaches may enhance retrieval precision, they often lack transparency and domain alignment. In contrast, the proposed GraphRAG system leverages structured knowledge from a Neo4j-based graph to ground retrieval and improve interpretability. The evaluation involved domain experts assessing the outputs of both systems, focusing on accuracy, relevance, and interpretability. Results: The GraphRAG system (System B) outperformed the hybrid vector-based baseline (System A) in tasks such as faulty component detection and identifying recurring issues. Expert feedback emphasized that System B provided better interpretability and stronger domain alignment. However, limitations were noted in cases where inputs were weak or task framing was ambiguous. Conclusions: Expert feedback confirms the value of structured knowledge integration, although task context and input quality can influence system performance. This work contributes to advancing knowledge-intensive domains, demonstrating that structured retrieval can enhance generative system outputs
Custodes Mutantur : The short term effect of mutation testing on software maintainability - a case study
No full textMutation testing is a technique for evaluating test effectiveness by introducing small faults into the code and measuring whether the test suite detects them. This thesis explores whether applying mutation testing in a targeted way has a measurable short-term impact on software maintainability and code quality. A case study was conducted in an industrial TypeScript codebase, where randomly selected files were subjected to thorough unit testing aimed at maximizing mutation scores. Maintainability metrics such as Maintainability Index, Halstead Metrics, and Cyclomatic Complexity were tracked before and after the intervention. The results showed modest improvements in several metrics, including a statistically significant correlation between increases in mutation score and maintainability index when outliers were excluded. However, no defects were uncovered, and many of the changes observed were not found to be statistically significant. While mutation testing did not result in dramatic short-term gains, it surfaced useful refactoring opportunities and reinforced earlier findings that mutation-guided testing can enhance code structure and testability. The study concludes that mutation testing holds value beyond metrics as a catalyst for cleaner, more maintainable code
Spelifiering för lärande : Multimodal applikation för bokläsning
No full textDen snabba digitala utvecklingen har möjliggjort nya former av lärande där spelifiering har blivit alltmer relevant metod för att öka engagemang. Denna studie undersöker hur en multimodal spelifierad applikation, Read&Focus, kan främja lärande genom bokläsning. Med utgångspunkt i tidigare forskning om motivationsteorier och spelifieringens designprinciper har en prototyp utvecklats och testats på en målgrupp. Analysen genomförs med hjälp av både kvantitativa och kvalitativa metoder och fokuserar på spelelements inverkan på användarens motivation, engagemang och lärande upplevelse. Resultaten visar att spelifiering genom spelelement såsom poängsystem, nivåer och interaktiva utmaningar kan förstärka både intresset för läsning och kunskapsinhämtning. Samtidigt identifieras utmaningar som kortvarig motivation och risk för ytligt lärande om designen inte är välbalanserad. Studien belyser även vikten av ett inkluderande perspektiv för att tillgodose olika användares behov.The rapid digital development has enabled new forms of learning where gamification has become an increasingly relevant method to enhance motivation and engagement. This study explores how a multimodal gamified application, Read&Focus, can promote learning throughbook reading. Building on previous research in motivation theories and gamification design principles, a prototype was developed and tested with group of users. The analysis combines qualitative and qualitative methods, focusing on how game elements influce user motivation, engagement, and the learning experience. The results indicate that gamification through features like point system, levels, and interactive challenges can enhance both reading interest and knowledge acquisition. However, challenges such as short-term motivation and the risk of superficial learning if poorly designed are also identified. The study highlights the importance of an inclusive approach to meet the diverse need of learners
Sustainability Assessment Tools for Small Scale Agri-Food Projects in Sweden
No full textThis study explored the usefulness of sustainability assessment tools for small-scale agri-food projects in Sweden, focusing on the 2000 m² project initiatives in Älvsbyn and Stockholm. Employing a critical, integrative approach, two frameworks that is the Five-Level Model of the Framework for Strategic Sustainable Development (FSSD) and Binder et al.’s (2010) framework guided the evaluation of the five selected tools: SMART, RISE, SAFE, SAEMETH, and LiteFarm. This research developed a conceptual framework combining these two methodologies, enabling a structured assessment across normative, systemic, and procedural dimensions. Results highlight SMART as the most comprehensive tool, offering broad coverage of environmental, social, and economic dimensions and robust procedural strengths. RISE follows. LiteFarm can also be used for its fair sustainability coverage, open access, and farmerfriendly visuals. SAFE and SAEMETH, despite their systemic and participatory design, show limitations in practical use and clarity. The findings suggest that SMART and RISE are best suited for full sustainability evaluations in complex contexts like the 2000 m² concept, while LiteFarm serves well for rapid assessments
Design and Development of Durable Conveyor Rollers for Outdoor Use
No full textBackground For industries utilizing conveyor systems, roller failure can result in reduced efficiency or even unplanned downtime. A typical conveyor roller consists of an outer tube, end caps at both ends, bearings housed within the end caps, and a throughgoing axle. In collaboration with NKT, this thesis investigates how outdoor conveyor rollers can be improved with respect to corrosion resistance, simplified maintenance, reduced weight, and sufficient structural performance, while also ensuring economic feasibility, manufacturability, and environmental sustainability. The project explores how rollers can be designed to perform reliably in outdoor environments under varying loads and how such designs can be optimized for large-scale manufacturing. Methods The work applied a Design Thinking approach to identify user needs and underlying challenges, generate new ideas, and explore these in a structured, iterative manner. Concepts were evaluated and refined through sketches, manual calculations, and primarily through Computer-Aided Design (CAD) modeling and numerical simulations. Results Several design alternatives were evaluated to improve the performance and reduce the weight of conveyor rollers. The most promising concepts demonstrated potential for extended service life, improved corrosion resistance, and weight reduction without compromising structural integrity. Some design ideas were discarded due to performance limitations. Conclusions The study concludes that it is feasible to develop a lighter and more cost-effective roller design with a potentially longer service life. This can be achieved by retaining the original tube while redesigning internal components such as the axle, bearings, and end caps to improve corrosion resistance and manufacturing efficiency.Bakgrund För industrier som använder transportband kan driftstopp eller minskad effektivitet uppstå om transportbandsrullar slutar fungera som de ska. En typisk transportbandsrulle består av en yttre tub, ändlock (end caps) i ytterändarna, lager inuti ändlocken samt en genomgående axel. I samarbete med NKT har detta examensarbete undersökt hur transportbandsrullar som används utomhus kan förbättras med avseende på korrosionsbeständighet, förenklat underhåll, viktminskning och bibehållen strukturell prestanda. Fokus har även legat på kostnadseffektivitet, tillverkningsbarhet och miljömässig hållbarhet. Arbetet behandlar hur rullarna kan designas för att fungera på ett tillförlitligt sätt i utomhusmiljö med varierande belastningar och hur konstruktionen kan optimeras för storskalig tillverkning. Metod En Design Thinking-process har tillämpats för att identifiera behov och bakomliggande problem, generera nya idéer samt för att strukturerat och iterativt utforska, förkasta eller vidareutveckla dessa idéer. Ritningar, beräkningar och numeriska simuleringar har genomförts både manuellt och med hjälp av datorstödd konstruktion (CAD), där CAD använts som huvudsakligt verktyg. Resultat Flera alternativa konstruktioner testades i syfte att minska vikt och förbättra hållbarhet hos rullens komponenter. Resultaten visar att vissa koncept kan möjliggöra viktminskning och ökad livslängd utan att kompromissa med hållfasthet eller funktion. Förbättringar identifierades även vad gäller korrosionsbeständighet och materialeffektivitet, medan andra idéer förkastades på grund av otillräcklig prestanda. Slutsats Det är möjligt att ta fram en lättare och mer kostnadseffektiv löpbandsrulle med förväntat längre livslängd. Detta kan uppnås genom att behålla originaltuben och samtidigt omkonstruera komponenter såsom axel, lager och ändlock för att förbättra korrosionsbeständighet och tillverkningsbarhet
Detection of AI-Generated Phishing Emails : Comparing The Efficiency Of SVM, Random Forest, CNN And BiLSTM In Detecting AI-Generated Phishing Emails
No full textBackground: Phishing attacks have become progressively more complex with the introduction of Large Language Models (LLMs) that are able to produce human-like emails. Consequently, automated phishing emails are difficult to differentiate from legitimate emails using conventional security measures. Accordingly, a new level of sophistication around phishing attacks is pushing for a next generation of detection mechanisms. Objectives: This thesis intends to detect and classify Artificial Intelligence (AI) generated emails from human-written phishing emails. Traditional machine learning models, Support Vector Machine (SVM) and Random Forest, are evaluated as well as DL models, Convolutional Neural Network (CNN) and Bidirectional Long Short Term Memory (BiLSTM). Methods: The study utilizes a multiclass email dataset from Kaggle and another spear-phishing dataset with AI-generated emails. Text preprocessing was done using natural language processing (NLP) processing steps, which included tokenizing and lemmatizing the text. The traditional machine learning models were executed using TF-IDF vectorized emails and the CNN and BiLSTM were executed using padded token sequences. In all cases the models were examined based on metrics of Accuracy, Precision, Recall, F1-score, and AUC-ROC. Results: All four models performed quite well, reaching perfect (100%) accuracy, precision, recall, F1-score, and AUC-ROC with SVM and BiLSTM. CNN and Random Forest were also effective, albeit with a few false positives. The study found a stark contrast in linguistic variables between AI-generated texts vs human-created texts, allowing us to classify the results accurately. Conclusions: Both traditional and DL models perform effectively in identifying AI generated phishing emails, with SVM and BiLSTM demonstrating the best performance. As this research illustrates, automated detection systems will be a practical tool in the arsenal of modern commonplace phishing defense. It is important however, to remain agile to evolutionary progression of AI.