Blekinge Institute of Technology
Not a member yet
13576 research outputs found
Sort by
Retrieval-Augmented Validation of AUTOSAR ARXML : A Machine-Learning-Assisted Code-Review Framework
No full textThe automotive industry heavily depends on the AUTOSAR (Automotive Open System Architecture) platform to standardize and streamline the development of electronic control units (ECUs). ARXML files, which serve as the structural backbone forAUTOSAR-based systems, pose unique challenges due to their hierarchical complexity and verbose nature. Manual review of these files is time-consuming, error-prone,and inefficient at the scale demanded by modern automotive workflows. Advances in artificial intelligence, particularly Retrieval-Augmented Generation (RAG) models, present an opportunity to automate and enhance ARXML validation. This thesis explores how integrating AUTOSAR specifications, vector embeddings, and promptengineering can improve the precision and efficiency of ARXML review processes. This research aims to design, implement, and evaluate a RAG-based pipeline tailored for ARXML analysis and review. The main objective is to assess whether RAGmodels can detect inconsistencies in ARXML configurations and deliver actionable,specification-grounded suggestions. An additional goal is to optimize developer experience by embedding AUTOSAR compliance logic directly into review feedback, thereby streamlining validation tasks. An iterative action research methodology was adopted to co-develop the systemwith practitioners. The framework integrates semantic vector embeddings, hybrid retrieval (BM25 + FAISS), and structured prompt templates to process ARXML diffs sand retrieve relevant segments from the AUTOSAR Classic Platform R23-11 specification. The system is evaluated using an LLM-as-a-Judge framework, which separately scores retrieval quality and generation faithfulness using rubric-based criteria. Performance metrics and model interpretability were analyzed across representative patch sets from a production-grade OEM project.The RAG system produced feedback that aligned well with AUTOSAR semantics, ordering clear and structured guidance to reviewers. The generation component received high ratings for clarity and completeness, although retrieval sometimeslacked clause-level grounding. While the tool reduced reliance on manual documentation lookup and improved review usability, it did not produce executable artefacts. The framework demonstrated interactive latency (~1.9s) suitable for integration withtools like Gerrit. This thesis demonstrates that a RAG-based approach can automate and improve ARXML review by embedding structured domain knowledge into AI feedback. Allthough the system currently targets AUTOSAR Classic R23-11, its modular architecture is adaptable to other structured formats such as YAML or JSON and applicable to regulated domains like aerospace and medical devices. Future work should address traceability, integration with live review systems, and broader generalization across standards and organizations
Securing Cloud-Native Applications (CNAs) : A Case Study of Practices in a large IT Company
No full textBackground: Cloud-native applications (CNAs), built using microservices and deployed via container orchestration platforms like Kubernetes, have significantly transformed how software is developed, deployed, and scaled. This shift from traditional monolithic architectures introduces new complexities in the realm of security. Unlike conventional IT environments, CNAs require a fundamentally different approach to securing code, infrastructure, and runtime environments due to their distributed, dynamic nature. Objectives: This thesis aims to explore the security practices employed in managing CNAs within a large IT organization. Specifically, it investigates how security responsibilities are shared across development, operations, and security teams using the 4Cs security framework — Code, Container, Cluster, and Cloud. The study further seeks to evaluate the effectiveness of modern DevSecOps practices, tools, and processes, while identifying ongoing security challenges faced by organizations operating in a cloud-native landscape. Method: The research adopts a case study methodology grounded in software engineering practices. Data collection was carried out through semi-structured interviews with DevSecOps professionals and a systematic review of organizations’ archival documents. To contextualize and validate these findings, a supplementary review of grey literature and industry security standards, such as OWASP, CIS Benchmarks, and DISA STIGs, was conducted to highlight common practices, gaps, and divergences in industry implementations. Results: The study reveals that while there is widespread adoption of automation, shift-left security, policy-as-code, and runtime monitoring, their effectiveness is often moderated by organizational culture and the level of ongoing training. Key challenges include balancing development speed with security rigor, managing alert fatigue, and responding to constantly evolving threats and misconfigurations. The research concludes with actionable recommendations that emphasize the importance of adaptive security strategies, continuous alignment with emerging standards, and the role of cross-functional collaboration in securing cloud-native environments
The Effects of Cybersecurity Training For Software Development Organizations
No full textBackground: This study examines the perspectives of developers and managers on software development organizations regarding the effectiveness of cyber security training in their everyday coding procedures. Purpose: The purpose of this study is to investigate how management’s expectations of up‑skilling and re‑skilling through cybersecurity‑awareness training, align with the skill gains that software developers themselves perceive. The research questions addressed in the study are: How do software engineers perceive cybersecurity training in relation to up-skilling and re-skilling? How does an organization's management perceive cybersecurity training? Method: A qualitative approach using semi-structured interviews with 6 managers and 8 developers from three different organizations were conducted. The selected companies are from Sweden and Greece, with one public sector agency and two private sector companies. The method of choice was based on the theory of innovation diffusion by Rogers. Empirical findings and analysis: From the analysis of all five characteristics of the innovation diffusion model, we see indications of high alignment between management and system development contexts on how cybersecurity training contributes to both up-skilling and re-skilling. This is especially observed in the view of compatibility, complexity, and relative advantage. However, partial alignment and misalignment were found in the remaining characteristics: observability, trialability. Conclusions: The study suggests that while developers experience cybersecurity training as practical and growth-oriented, management requires clearer evidence of return and observable impact to support sustained investment. It is observed that management is lacking pre-investment analysis. Recommendations for future research: We propose for future research to focus on comprehensive data that includes more companies within the sector. Furthermore, combinations of quantitative and qualitative methods to extract useful data that helps in remediating the various misalignments between developers and managers could be conducted
Urban resiliens i praktiken : En kvalitativ analys av tre kommuners resiliensstrategier
No full textAs climate change becomes increasingly evident, there is a growing need for cities to be equipped to meet its challenges, such as flooding from intense rainfall, rising sea levels, and heatwaves. This study explores how three Swedish municipalities, Malmö, Gävle, and Norrköping, work with climate adaptation in the context of physical planning, with a particular focus on how these efforts promote urban resilience. The study is conducted as a qualitative multiple-case study and is based on document analysis and semi-structured interviews with municipal officials. The analysis is grounded in a theoretical framework centered on five key qualities of urban resilience: preparedness, redundancy, robustness, flexibility, and innovation. The results show that all three municipalities have implemented extensive climate adaptation measures, but their focus and strategies differ. For example, Malmö has developed a comprehensive strategy to protect the city from sea level rise through technical infrastructure and coastal defenses. Gävle has prioritized financial investments and nature-based solutions, while Norrköping focuses more on internal coordination, vulnerability assessments, and addressing particularly exposed areas. The study also highlights that lessons learned from previous climate-related events, such as floods and heatwaves, have acted as driving forces in the development of more resilient urban planning. Through the analysis, the specific measures and strategies adopted by each municipality are made concrete. Furthermore, the analysis provides valuable insights into how physical planning can contribute to increased resilience in the urban environment. The conclusion is that climate adaptation efforts play a crucial role in strengthening urban resilience. A successful adaptation process requires the integration of the five urban resilience qualities identified in the theoretical framework. However, the study also shows that there are different approaches to operationalizing these qualities. The experiences developed in Malmö, Gävle, and Norrköping may serve as guidance for other municipalities, both in Sweden and globally, as cities around the world are likely to face similar climate-related challenges. I takt med att klimatförändringarna blir alltmer påtagliga så finns det ett ökat behov av att städer rustas för att kunna möta dess utmaningar, det kan exempelvis röra sig om översvämningar från skyfall eller stigande havsnivåer och värmeböljor. Denna studie undersöker hur tre svenska kommuner, Malmö, Gävle och Norrköping, arbetar med klimatanpassning inom den fysiska planeringen, med särskilt fokus på hur dessa åtgärder främjar urban resiliens. Studien har genomförts som en kvalitativ flerfallsstudie och bygger på dokumentanalyser och semistrukturerade intervjuer med kommunala tjänstepersoner som datainsamling. Analysen utgår från ett teoretiskt ramverk baserat på fem kvaliteter inom urban resiliens: förberedelse, reservkapacitet, robusthet, flexibilitet och innovation. Resultaten visar att samtliga kommuner har vidtagit omfattande klimatanpassningsåtgärder, men att deras fokus och strategier skiljer sig åt. Malmö har exempelvis en starkt utbyggd strategi för att skydda staden mot stigande havsnivåer genom teknisk infrastruktur och kustskydd, medan Gävle har prioriterat ekonomiska satsningar och naturbaserade lösningar. Norrköping arbetar mer med intern samordning, sårbarhetsanalyser och särskilt utsatta områden. Studien visar också att lärdomar från tidigare klimatrelaterade händelser, såsom översvämningar och värmeböljor, har fungerat som en drivande faktor för att utveckla mer resilient stadsplanering. Genom analysen så konkretiseras de åtgärder och strategier som respektive kommun valt att utveckla. Analysen bidrar vidare till viktiga insikter om hur den fysiska planeringen kan bidra till ökad resiliens i den urbana miljön. Slutsatsen är att kommunernas klimatanpassningsarbete har stor betydelse för att stärka den urbana resiliensen, och att en framgångsrik anpassning förutsätter de fem urban resilienta kvaliteter det teoretiska ramverket är uppbyggt på, men att det däremot finns olika tillvägagångssätt inom dessa kvaliteter. De erfarenheter som Malmö, Gävle och Norrköping har utvecklat kan fungera som vägledning för andra kommuner både i Sverige, men också andra städer i världen som med stor sannolikhet kommer att möta liknande klimatrelaterade utmaningar.
To see and hear paintings : How can gaze-guided interactive sound bring two-dimensional art to life?
No full textDen tekniska utvecklingen går ständigt framåt och nyare konstformer har börjat konkurrera med de tvådimensionella statiska visuella medierna mer och mer. Skulle en introduktion av interaktivitet och ett nytt sinne kunna ge ett större intresse för denna konst? Vi ville med denna undersökningen kolla på hur en kan utveckla upplevelsen för att konsumera stilla konst, och på så sätt göra det mer intressant med adderingen av ett sinne. För att ta reda på detta skapades en gestaltning där en ljudbild förändrades beroende på var användaren tittade, med hjälp av en eye tracker. Denna gestaltningen kom att bli den huvudsakliga komponenten för denna undersökning. Denna prototyp användes sedan för att utföra användartester där deltagarna fick testa gestaltningen och besvara frågor kring deras upplevelse. Dessa användartester låg sedan till grund för undersökningens resultat och leder oss till svaret på vår forskningsfråga. Resultatet på undersökningen visar att tekniken inte bara förhöjer upplevelsen genom att göra denna mer stimulerande, men ökar även intresset för konsten.Technological development is constantly advancing and newer art forms have begun to compete with the two-dimensional static visual media more. Could the introduction of interactivity and a new sense provide greater interest in this art? With this study, we wanted to look at how one can develop the experience of consuming still art, and thus make it more interesting with the addition of a sense. To find this out, a design was created where a soundscape changed depending on the user’s gaze, using an eye tracker. This design became the main component of this study. This prototype was then used to carry out user tests where participants tested the design and answer questions about their experience. These user tests then formed the basis for the results of the study and led us to the answer to our research question. The results of the study show that this technology not only enhances the experience by making it more stimulating, but also increases the interest in the art
Shifting Gears in the Automotive Industry: The Role of EU Sustainability Regulations in Business Model Innovation : (A Comparative case study of Volvo Group and Lynk&Co International)
No full textPurpose - This study investigates the intricate relationship between EU regulatory frameworks and business model innovation (BMI) in the automotive sector, with a particular focus on how sustainability imperatives and evolving market dynamics shape innovation strategies. Starting from exploring if regulations accelerate or impede innovation, to further drill down on how and what strategic approaches do companies adopt? Method – A mixed method approach strategically combining qualitative and quantitative research design was adopted using data collected from semi-structured interviews, industry-wide survey and publicly available reports from the case companies: Volvo Group and Lynk&Co International. This approach enabled in-depth exploration of the influence of sustainability regulations on strategic innovation efforts in automotive business. Analysis/Findings – An Iterative Analysis Model using the Gioia methodology was used, where the results from interviews and surveys were categorized into concepts. Related concepts were then grouped into aggregate dimensions and finally a framework was developed. This framework shows the dynamic interplay of the concepts and dimensions. Key Findings suggest that EU Sustainability Regulations are mostly accelerating BMI with specific challenges and BMI potentials. With certain Strategic Adaptations and right Systemic support, automotive firms can balance compliance and innovation. Recent advancements point towards a shift from reactive compliance to proactive innovation practices to withstand market competition and sustain brand positioning. Limitations and Future Research - Given its geographic and contextual focus, the study’s findings may not be universally applicable. Comparative research across diverse regions, such as North America, Asia, and Latin America could provide broader insights. Additionally, future studies should examine how differing regulatory, cultural, and economic contexts influence innovation pathways.
An Application of Crystal Plasticity to predict Forming Limit Curve of Dual-Phase Steels with Validation
No full textIn this paper an alternative approach for generating Forming Limit Curve (FLC) of a Dual-Phase (DP800) steel is presented and validated. Investigation was performed using a rate-dependent crystal plasticity (CP) framework solved through a spectral solver and combined subsequently with the Marciniak-Kuczynski (M-K) approach for the evaluation of FLC using the commercial software Digimat. Representative Volume Element (RVE) was constructed using the built-in Voronoi tessellation method within Digimat from the Electron Back Scattering Diffraction (EBSD) measurements on the material. The hardening parameters were calibrated through inverse optimisation to match the experimental uniaxial tensile behaviour of the material. The calibrated model was then used as a virtual testing tool to predict the anisotropic yield behaviour of the DP800 steel and the Yld2000 function was fitted to the predicted anisotropic yield surface. The fitted Yld2000 function was then used to determine the FLC using the M-K approach implemented in Digimat where the sheet necking is modelled through an initial imperfection as a narrow band with reduced thickness. The whole workflow was successfully validated with experimental evaluation of FLC. The results are satisfactory and demonstrate the suitability of the presented workflow for FLC evaluation of advanced steels
Player Behaviour in a Creative Sandbox Game : A Comparative Study of Player Behaviour Regarding Computer, Phone, and Game Proficiency in Creative Sandbox Games
No full textIntrinsic motivation is an underappreciated aspect when discussing the design and creation of videogames. In today’s landscape, it is much more common to use extrinsic motivators when designing video games as they rely less on player behaviour and can be easier to predict. However, as some of the most popular and influential games make heavy use of intrinsic motivators by giving a high degree of player freedom, an argument can be made that more research and discussion should be made for game designers to consider using intrinsic motivators when designing video games. This thesis’s aim is to gather data from participants engaging in a Creative Video Game study in an attempt to answer whether player behaviour can be traced from demographic data. The game is a 3D first person perspective building game which lets player place and destroy objects on a grid-based map. The study uses behavioural themes to categorize the different demographics for analysis and comparisons; this allowed for an easy-to-understand evaluation of whether player behaviour was different between demographics. Furthermore, the study also takes the opportunity to analyse how different point of interest structures affect a player’s decision-making and behaviour in a creative game.The results show that behavioural differences appeared between the different demographic groups through the analysis of 3D plotted graphs and raw data bar graphs as well as through the assignment of the different behavioural themes. The results indicate that demographic patterns could exists but that further research using a larger and more detailed study would be needed. The results in regard to the structures and their effect on player decision led to interesting results. The structures affected the way players moved around the game environment, yet they were not interesting enough to fully capture a player’s attention. Even though the structures failed to fully draw player attention in regard to placing and destroying objects, a clear favourite structure did emerge with its properties being ease of access, incompleteness, and familiarity
Vad orsakar plockfel? En kvantitativ analys av faktorer som påverkar förekomsten och omfattningen av plockfel : En studie av felorsaker i varulager med hänsyn till plockfel, dess förekomst och omfattning
No full textBackground: Order picking is the process of retrieving articles from storage locations to fulfill an order from a customer. When the warehouse fails to fulfill the customer’s order, it is referred to as a picking error. Most of the order picking is still done manually by operators. Manual order picking is the most cost-intensive operation of the warehouse. This is because picking errors result in expensive reclamations due to additional administrative work and picking to correct the order. However, the factors behind picking errors in order picking have received limited attention in previous research. Objectives: The purpose of this study is to generate new knowledge regarding which human, organizational, and technical factors influence the occurrence of picking errors in warehouse operations. Methods: This study is based on a quantitative research design centered on two different regression models, a Firth’s penalized logistic regression and a truncated negative binomial regression. The empirical data has been collected from a warehouse in Sweden for ten months with affiliated order reclamations. Results: The result indicated that picking errors are associated with factors linked to physical, mental, and organizational aspects. Amongst others, the results showed that the number of articles in an order, picking time, picking zones, and the number of operators working on an order all influence the probability of an order having a picking error. With respect to the quantity of picking errors per order, the result indicated that factors such as total weight, picking zones, the amount of time until the next break, the number of unique articles in an order, and the number of operators working on an order all influence the size of the picking error. Conclusions: The study concludes that both organizational, human, and technological factors influence the occurrence and the quantity of a picking error. More specifically, factors related to mental and physical fatigue can be linked to the reason why these types of errors occur. Bakgrund: Orderplock definieras som processen att hämta artiklar från lagerplatser för att kunna fullfölja en order från en kund. När lagret misslyckas med att uppfylla kundens beställning kallas det för ett plockfel. Det mesta av orderplockningen sker fortfarande manuellt av operatörer. Manuell orderplockning är den mest kostnadskrävande processen i ett varulager. Detta är eftersom plockfel resulterar i dyra reklamationer på grund av ytterligare administrativt arbete och plockning för att rätta beställningen. Faktorerna bakom plockfel vid orderplock har dock inte uppmärksammats särskilt mycket i tidigare forskning. Syfte: Syftet med denna studie är att införskaffa ny kunskap om vilka mänskliga, organisatoriska och tekniska faktorer som påverkar uppkomsten av plockfel i lagerverksamheter. Metod: Denna studie är baserad på en kvantitativ forskningsdesign och omfattar två olika regressionsmodeller, en Firths straffade logistisk regression och en trunkerad negativ binomial regression. Empirin har samlats in från ett varulager i Sverige under tio månader med tillhörande orderreklamationer. Resultat: Resultatet indikerade att plockfel är förknippade med faktorer kopplade till fysiska, psykiska och organisatoriska aspekter. Bland annat visade resultatet att antalet artiklar i en order, plocktid, plockzoner och antalet operatörer som arbetar med en order, alla påverkar sannolikheten för att en order innehar ett plockfel. När det gäller antalet plockfel per order visade resultatet att faktorer som totalvikt, plockzoner, tid till nästa rast, antalet unika artiklar i en order och antalet operatörer som arbetar med en order, alla påverkar storleken på plockfelet. Slutsatser: Studien visar således att både organisatoriska, mänskliga och tekniska faktorer påverkar förekomsten och omfattningen av ett plockfel. Mer specifikt kan faktorer relaterade till mental och fysisk trötthet kopplas till orsaken till att dessa typer av fel uppstår.
From Academia to Professional Practice : Development of Key Competencies in Sustainability through the ABCD Process
No full textThis research explores how the ABCD Process, a strategic planning approach within the Framework for Strategic Sustainable Development, supports the development of key competencies in sustainability in professional practice. Addressing a persistent gap between academia and professional settings, the research adopts a qualitative methodology with semi-structured interviews with two leading experts and 14 practitioners from diverse sectors and experiences. Findings indicate that the ABCD Process facilitates the cultivation of all eight KCs with varying prominence and context-dependence. Planning Competencies emerge mostly in specific steps, e.g., the A-Step fosters Systems, Futures, and Values Thinking, while the D-Step supports Strategies Thinking. In contrast, Interpersonal, Intra-personal, and Integration Competencies develop across the whole process. Implementation Competency was least supported, often requiring additional steps for its activation. Practitioners highlighted the flexibility of the ABCD Process but also shared different views on recognizing competencies. Experts emphasized the process’ theoretical grounding in transdisciplinary, stakeholder-driven designs. Limitations include a deductive design that may constrain emergent themes, uneven familiarity with key competencies among interviewees, and absence of standardized competency assessment. Future research could explore co-designing ABCD Process adaptations with practitioners to enhance specific competencies and develop robust, context-sensitive evaluations